The prediction of molecular properties is a crucial task in the field of drug discovery. Computational methods that can accurately predict molecular properties can significantly accelerate the drug discovery process and reduce the cost of drug discovery. In recent years, iterative updates in computing hardware and the rise of deep learning have created a new and effective path for molecular property prediction. Deep learning methods can leverage the vast amount of data accumulated over the years in drug discovery and do not require complex feature engineering. In this review, we summarize molecular representations and commonly used datasets in molecular property prediction models and present advanced deep learning methods for molecular property prediction, including state‐of‐the‐art deep learning networks such as graph neural networks and Transformer‐based models, as well as state‐of‐the‐art deep learning strategies such as 3D pre‐train, contrastive learning, multi‐task learning, transfer learning, and meta‐learning. We also point out some critical issues such as lack of datasets, low information utilization, and lack of specificity for diseases.
预测分子性质是药物发现领域的一项重要任务。能够准确预测分子性质的计算方法可以大大加快药物发现的进程，降低药物发现的成本。近年来，计算硬件的迭代更新和深度学习的兴起为分子性质预测开辟了一条新的有效途径。深度学习方法可以利用药物发现过程中多年积累的大量数据，而且不需要复杂的特征工程。在这篇综述中，我们总结了分子性质预测模型中的分子表征和常用数据集，并介绍了用于分子性质预测的先进深度学习方法，包括最先进的深度学习网络（如图神经网络和基于 Transformer 的模型），以及最先进的深度学习策略（如 3D 预训练、对比学习、多任务学习、迁移学习和元学习）。我们还指出了一些关键问题，如缺乏数据集、信息利用率低、缺乏疾病特异性等。
{"title":"Advanced deep learning methods for molecular property prediction","authors":"Chao Pang, Henry H. Y. Tong, Leyi Wei","doi":"10.1002/qub2.23","DOIUrl":"https://doi.org/10.1002/qub2.23","url":null,"abstract":"The prediction of molecular properties is a crucial task in the field of drug discovery. Computational methods that can accurately predict molecular properties can significantly accelerate the drug discovery process and reduce the cost of drug discovery. In recent years, iterative updates in computing hardware and the rise of deep learning have created a new and effective path for molecular property prediction. Deep learning methods can leverage the vast amount of data accumulated over the years in drug discovery and do not require complex feature engineering. In this review, we summarize molecular representations and commonly used datasets in molecular property prediction models and present advanced deep learning methods for molecular property prediction, including state‐of‐the‐art deep learning networks such as graph neural networks and Transformer‐based models, as well as state‐of‐the‐art deep learning strategies such as 3D pre‐train, contrastive learning, multi‐task learning, transfer learning, and meta‐learning. We also point out some critical issues such as lack of datasets, low information utilization, and lack of specificity for diseases.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"33 2","pages":""},"PeriodicalIF":3.1,"publicationDate":"2023-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139259366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Feiran Li, Yu Chen, Johan Gustafsson, Hao Wang, Yi Wang, Chong Zhang, Xinhui Xing
Abstract Over the last 15 years, genome‐scale metabolic models (GEMs) have been reconstructed for human and model animals, such as mouse and rat, to systematically understand metabolism, simulate multicellular or multi‐tissue interplay, understand human diseases, and guide cell factory design for biopharmaceutical protein production. Here, we describe how metabolic networks can be represented using stoichiometric matrices and well‐defined constraints for flux simulation. Then, we review the history of GEM development for quantitative understanding of Homo sapiens and other relevant animals, together with their applications. We describe how model develops from H . sapiens to other animals and from generic purpose to precise context‐specific simulation. The progress of GEMs for animals greatly expand our systematic understanding of metabolism in human and related animals. We discuss the difficulties and present perspectives on the GEM development and the quest to integrate more biological processes and omics data for future research and translation. We truly hope that this review can inspire new models developed for other mammalian organisms and generate new algorithms for integrating big data to conduct more in‐depth analysis to further make progress on human health and biopharmaceutical engineering.
{"title":"Genome‐scale metabolic models applied for human health and biopharmaceutical engineering","authors":"Feiran Li, Yu Chen, Johan Gustafsson, Hao Wang, Yi Wang, Chong Zhang, Xinhui Xing","doi":"10.1002/qub2.21","DOIUrl":"https://doi.org/10.1002/qub2.21","url":null,"abstract":"Abstract Over the last 15 years, genome‐scale metabolic models (GEMs) have been reconstructed for human and model animals, such as mouse and rat, to systematically understand metabolism, simulate multicellular or multi‐tissue interplay, understand human diseases, and guide cell factory design for biopharmaceutical protein production. Here, we describe how metabolic networks can be represented using stoichiometric matrices and well‐defined constraints for flux simulation. Then, we review the history of GEM development for quantitative understanding of Homo sapiens and other relevant animals, together with their applications. We describe how model develops from H . sapiens to other animals and from generic purpose to precise context‐specific simulation. The progress of GEMs for animals greatly expand our systematic understanding of metabolism in human and related animals. We discuss the difficulties and present perspectives on the GEM development and the quest to integrate more biological processes and omics data for future research and translation. We truly hope that this review can inspire new models developed for other mammalian organisms and generate new algorithms for integrating big data to conduct more in‐depth analysis to further make progress on human health and biopharmaceutical engineering.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"8 9","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136352058","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Recently, Quantitative Biology (QB) held a discussion on “AI (artificial intelligence) for Life Science” among editorial board members and interested scholars in anticipation of rapid development of this growing area after AlphaGo and ChatGPT mania. Many young people tend to get confused between facts and fictions; heated debates are unavoidable even among their mentors. When deep learning as represented by convolutional neural networks and LSTM (long short-term memory) was made available for bioinformatics students, many of them rushed into this research field and tried to adopt these methods in all their projects without knowing the history that these tools were becoming successful consistently with Moore’s Law (relating to rapid computer technology advances), but more importantly due to new structural/functional understanding of vision and auditory circuits in the brain. Recently, some young people have claimed “LSTM is dead, long live transformer” (which is somewhat like saying “the bike is dead, long live the car”), and have amplified the threat that ChatGPT could wipe out human jobs. They believe transformer is the “silver bullet” for all learning tasks, clearly reflecting their lack of basic knowledge (i.e. “No Free Lunch Theory,” the trade-off of such global “attention network” is to pay the price for complexity: difficulty of training and high memory costs). There is no doubt ML (machine learning) and AI have brought a new revolution in science and technology, and will deliver huge unforeseeable impact to human everyday life as well as to social relationships. In this context, QB journal could be a great platform for encouraging intellectual discussions and for promoting AI for Life Science. Here, I would like to use the DIALOG to “抛砖引玉” (make some initial remarks to get the ball rolling), although it is my personal opinion which is inevitably subject to bias and limitations. AI: Do you know my name “Artificial Intelligence” is defined by the Oxford English Dictionary as the capacity of computer systems (which may be referred as a “robot”) to exhibit or simulate your intelligent behavior? NI: Wait a minute, intelligence itself is defined as the ability to learn, understand and think in a logical way. Can you think? AI: No. But that definition is too restrictive, actually intelligence has different scopes and degrees. Simple intelligent control devices date back to antiquity, from windmills to thermostat. NI: Agree, everything is relative. Macromolecules (e.g., enzyme) and cells (e.g., immune cell) might be considered to be intelligent; see how a white blood cell is chasing bacteria in the youtube website (search for “Crawling neutrophil chasing a bacterium”). Our emergent/collective intelligent behavior does not require a brain or even a neuron; see how slime molds can solve optimization—Hamilton cycle-problem more effectively than a human in the youtube website (search for “Intelligence without a brain?”). Before there was any neuron, C
{"title":"Dialog between artificial intelligence & natural intelligence","authors":"Michael Q. Zhang","doi":"10.1002/qub2.5","DOIUrl":"https://doi.org/10.1002/qub2.5","url":null,"abstract":"Recently, Quantitative Biology (QB) held a discussion on “AI (artificial intelligence) for Life Science” among editorial board members and interested scholars in anticipation of rapid development of this growing area after AlphaGo and ChatGPT mania. Many young people tend to get confused between facts and fictions; heated debates are unavoidable even among their mentors. When deep learning as represented by convolutional neural networks and LSTM (long short-term memory) was made available for bioinformatics students, many of them rushed into this research field and tried to adopt these methods in all their projects without knowing the history that these tools were becoming successful consistently with Moore’s Law (relating to rapid computer technology advances), but more importantly due to new structural/functional understanding of vision and auditory circuits in the brain. Recently, some young people have claimed “LSTM is dead, long live transformer” (which is somewhat like saying “the bike is dead, long live the car”), and have amplified the threat that ChatGPT could wipe out human jobs. They believe transformer is the “silver bullet” for all learning tasks, clearly reflecting their lack of basic knowledge (i.e. “No Free Lunch Theory,” the trade-off of such global “attention network” is to pay the price for complexity: difficulty of training and high memory costs). There is no doubt ML (machine learning) and AI have brought a new revolution in science and technology, and will deliver huge unforeseeable impact to human everyday life as well as to social relationships. In this context, QB journal could be a great platform for encouraging intellectual discussions and for promoting AI for Life Science. Here, I would like to use the DIALOG to “抛砖引玉” (make some initial remarks to get the ball rolling), although it is my personal opinion which is inevitably subject to bias and limitations. AI: Do you know my name “Artificial Intelligence” is defined by the Oxford English Dictionary as the capacity of computer systems (which may be referred as a “robot”) to exhibit or simulate your intelligent behavior? NI: Wait a minute, intelligence itself is defined as the ability to learn, understand and think in a logical way. Can you think? AI: No. But that definition is too restrictive, actually intelligence has different scopes and degrees. Simple intelligent control devices date back to antiquity, from windmills to thermostat. NI: Agree, everything is relative. Macromolecules (e.g., enzyme) and cells (e.g., immune cell) might be considered to be intelligent; see how a white blood cell is chasing bacteria in the youtube website (search for “Crawling neutrophil chasing a bacterium”). Our emergent/collective intelligent behavior does not require a brain or even a neuron; see how slime molds can solve optimization—Hamilton cycle-problem more effectively than a human in the youtube website (search for “Intelligence without a brain?”). Before there was any neuron, C","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"241 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135974128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Simulating the whole brain as an alternative way to achieve AGI","authors":"Jianfeng Feng","doi":"10.1002/qub2.6","DOIUrl":"https://doi.org/10.1002/qub2.6","url":null,"abstract":"","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"76 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136104454","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The flourishing plant science promotes the exploding number of data and the expansion of toolkits. Plant synthetic biology is still in its early stages and requires more quantitative and predictable study. Despite the challenges, some pioneering examples have been successfully demonstrated in model plants. Backgrounds As an increasing number of synthetic switches and circuits have been created for plant systems and of synthetic products produced in plant chassis, plant synthetic biology is taking a strong foothold in agriculture and medicine. The ever‐exploding data has also promoted the expansion of toolkits in this field. Genetic parts libraries and quantitative characterization approaches have been developed. However, plant synthetic biology is still in its infancy. The considerations for selecting biological parts to design and construct genetic circuits with predictable functions remain desired. Results In this article, we review the current biotechnological progresses in field of plant synthetic biology. Assembly standardization and quantitative approaches of genetic parts and genetic circuits are discussed. We also highlight the main challenges in the iterative cycles of design‐build‐test‐learn for introducing novel traits into plants. Conclusion Plant synthetic biology promises to provide important solutions to many issues in agricultural production, human health care, and environmental sustainability. However, tremendous challenges exist in this field. For example, the quantitative characterization of genetic parts is limited; the orthogonality and the transfer functions of circuits are unpredictable; and also, the mathematical modeling‐assisted circuits design still needs to improve predictability and reliability. These challenges are expected to be resolved in the near future as interests in this field are intensifying.
{"title":"From qualitative to quantitative: the state of the art and challenges for plant synthetic biology","authors":"Chenfei Tian, Jianhua Li, Yong Wang","doi":"10.15302/j-qb-022-0326","DOIUrl":"https://doi.org/10.15302/j-qb-022-0326","url":null,"abstract":"The flourishing plant science promotes the exploding number of data and the expansion of toolkits. Plant synthetic biology is still in its early stages and requires more quantitative and predictable study. Despite the challenges, some pioneering examples have been successfully demonstrated in model plants. Backgrounds As an increasing number of synthetic switches and circuits have been created for plant systems and of synthetic products produced in plant chassis, plant synthetic biology is taking a strong foothold in agriculture and medicine. The ever‐exploding data has also promoted the expansion of toolkits in this field. Genetic parts libraries and quantitative characterization approaches have been developed. However, plant synthetic biology is still in its infancy. The considerations for selecting biological parts to design and construct genetic circuits with predictable functions remain desired. Results In this article, we review the current biotechnological progresses in field of plant synthetic biology. Assembly standardization and quantitative approaches of genetic parts and genetic circuits are discussed. We also highlight the main challenges in the iterative cycles of design‐build‐test‐learn for introducing novel traits into plants. Conclusion Plant synthetic biology promises to provide important solutions to many issues in agricultural production, human health care, and environmental sustainability. However, tremendous challenges exist in this field. For example, the quantitative characterization of genetic parts is limited; the orthogonality and the transfer functions of circuits are unpredictable; and also, the mathematical modeling‐assisted circuits design still needs to improve predictability and reliability. These challenges are expected to be resolved in the near future as interests in this field are intensifying.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135735878","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Computational methods for DDIs and DTIs prediction are essential for accelerating the drug discovery process. We proposed a novel deep learning method DeepDrug, to tackle these two problems within a unified framework. DeepDrug is capable of extracting comprehensive features of both drug and target protein, thus demonstrating a superior prediction performance in a series of experiments. The downstream applications show that DeepDrug is useful in facilitating drug repositioning and discovering the potential drug against specific disease. Background Computational approaches for accurate prediction of drug interactions, such as drug‐drug interactions (DDIs) and drug‐target interactions (DTIs), are highly demanded for biochemical researchers. Despite the fact that many methods have been proposed and developed to predict DDIs and DTIs respectively, their success is still limited due to a lack of systematic evaluation of the intrinsic properties embedded in the corresponding chemical structure. Methods In this paper, we develop DeepDrug, a deep learning framework for overcoming the above limitation by using residual graph convolutional networks (Res‐GCNs) and convolutional networks (CNNs) to learn the comprehensive structure‐ and sequence‐based representations of drugs and proteins. Results DeepDrug outperforms state‐of‐the‐art methods in a series of systematic experiments, including binary‐class DDIs, multi‐class/multi‐label DDIs, binary‐class DTIs classification and DTIs regression tasks. Furthermore, we visualize the structural features learned by DeepDrug Res‐GCN module, which displays compatible and accordant patterns in chemical properties and drug categories, providing additional evidence to support the strong predictive power of DeepDrug. Ultimately, we apply DeepDrug to perform drug repositioning on the whole DrugBank database to discover the potential drug candidates against SARS‐CoV‐2, where 7 out of 10 top‐ranked drugs are reported to be repurposed to potentially treat coronavirus disease 2019 (COVID‐19). Conclusions To sum up, we believe that DeepDrug is an efficient tool in accurate prediction of DDIs and DTIs and provides a promising insight in understanding the underlying mechanism of these biochemical relations.
{"title":"DeepDrug: A general graph-based deep learning framework for drug-drug interactions and drug-target interactions prediction","authors":"Qijin Yin, Rui Fan, Xusheng Cao, Qiao Liu, Rui Jiang, Wanwen Zeng","doi":"10.15302/j-qb-022-0320","DOIUrl":"https://doi.org/10.15302/j-qb-022-0320","url":null,"abstract":"Computational methods for DDIs and DTIs prediction are essential for accelerating the drug discovery process. We proposed a novel deep learning method DeepDrug, to tackle these two problems within a unified framework. DeepDrug is capable of extracting comprehensive features of both drug and target protein, thus demonstrating a superior prediction performance in a series of experiments. The downstream applications show that DeepDrug is useful in facilitating drug repositioning and discovering the potential drug against specific disease. Background Computational approaches for accurate prediction of drug interactions, such as drug‐drug interactions (DDIs) and drug‐target interactions (DTIs), are highly demanded for biochemical researchers. Despite the fact that many methods have been proposed and developed to predict DDIs and DTIs respectively, their success is still limited due to a lack of systematic evaluation of the intrinsic properties embedded in the corresponding chemical structure. Methods In this paper, we develop DeepDrug, a deep learning framework for overcoming the above limitation by using residual graph convolutional networks (Res‐GCNs) and convolutional networks (CNNs) to learn the comprehensive structure‐ and sequence‐based representations of drugs and proteins. Results DeepDrug outperforms state‐of‐the‐art methods in a series of systematic experiments, including binary‐class DDIs, multi‐class/multi‐label DDIs, binary‐class DTIs classification and DTIs regression tasks. Furthermore, we visualize the structural features learned by DeepDrug Res‐GCN module, which displays compatible and accordant patterns in chemical properties and drug categories, providing additional evidence to support the strong predictive power of DeepDrug. Ultimately, we apply DeepDrug to perform drug repositioning on the whole DrugBank database to discover the potential drug candidates against SARS‐CoV‐2, where 7 out of 10 top‐ranked drugs are reported to be repurposed to potentially treat coronavirus disease 2019 (COVID‐19). Conclusions To sum up, we believe that DeepDrug is an efficient tool in accurate prediction of DDIs and DTIs and provides a promising insight in understanding the underlying mechanism of these biochemical relations.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135944512","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Morphogenesis is a complex process in a developing animal at the organ, cellular and molecular levels. In this investigation, allometry at the cellular level was evaluated.Geometric information, including the time‐lapse Cartesian coordinates of each cell’s center, was used for calculating the allometric coefficients. A zero‐centroaxial skew‐symmetrical matrix ( CSSM), was generated and used for constructing another square matrix (basic square matrix: BSM), then the determinant of BSM was calculated ( d). The logarithms of absolute d (Lad) of cell group at different stages of development were plotted for all of the cells in a range of development stages; the slope of the regression line was estimated then used as the allometric coefficient. Moreover, the lineage growth rate (LGR) was also calculated by plotting the Lad against the logarithm of the time. The complexity index at each stage was calculated. The method was tested on a developing Caenorhabditis elegans embryo.We explored two out of the four first generated blastomeres in C. elegans embryo. The ABp and EMS lineages show that the allometric coefficient of ABp was higher than that of EMS, which was consistent with the complexity index as well as LGR.The conclusion of this study is that the complexity of the differentiating cells in a developing embryo can be evaluated by allometric scaling based on the data derived from the Cartesian coordinates of the cells at different stages of development.
{"title":"Cell‐based allometry: an approach for evaluation of complexity in morphogenesis","authors":"Ali Tarihi, Mojtaba Tarihi, T. Tiraihi","doi":"10.15302/j-qb-022-0319","DOIUrl":"https://doi.org/10.15302/j-qb-022-0319","url":null,"abstract":"Morphogenesis is a complex process in a developing animal at the organ, cellular and molecular levels. In this investigation, allometry at the cellular level was evaluated.Geometric information, including the time‐lapse Cartesian coordinates of each cell’s center, was used for calculating the allometric coefficients. A zero‐centroaxial skew‐symmetrical matrix ( CSSM), was generated and used for constructing another square matrix (basic square matrix: BSM), then the determinant of BSM was calculated ( d). The logarithms of absolute d (Lad) of cell group at different stages of development were plotted for all of the cells in a range of development stages; the slope of the regression line was estimated then used as the allometric coefficient. Moreover, the lineage growth rate (LGR) was also calculated by plotting the Lad against the logarithm of the time. The complexity index at each stage was calculated. The method was tested on a developing Caenorhabditis elegans embryo.We explored two out of the four first generated blastomeres in C. elegans embryo. The ABp and EMS lineages show that the allometric coefficient of ABp was higher than that of EMS, which was consistent with the complexity index as well as LGR.The conclusion of this study is that the complexity of the differentiating cells in a developing embryo can be evaluated by allometric scaling based on the data derived from the Cartesian coordinates of the cells at different stages of development.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"18 1","pages":""},"PeriodicalIF":3.1,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139371881","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Use of artificial neural networks to identify and analyze polymerized actin-based cytoskeletal structures in 3D confocal images","authors":"","doi":"10.15302/j-qb-022-0325","DOIUrl":"https://doi.org/10.15302/j-qb-022-0325","url":null,"abstract":"","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"1 1","pages":""},"PeriodicalIF":3.1,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67351718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yuanpeng Xiong, Xuan He, Dan Zhao, Tao Jiang, Jianyang Zeng
Background : Chromatin-associated RNA (caRNA) acts as a ubiquitous epigenetic layer in eukaryotes, and has been reported to be essential in various biological processes, including gene transcription, chromatin remodeling and cellular differentiation. Recently, numerous experimental techniques have been developed to characterize genome-wide RNA-chromatin interactions to understand their underlying biological functions. However, these experimental methods are generally expensive, time-consuming, and limited in identifying all potential sites, while most of the existing computational methods are restricted to detecting only specific types of RNAs interacting with chromatin. Methods : Here, we propose a highly interpretable computational framework, named DeepRCI, to identify the interactions between various types of RNAs and chromatin. In this framework, we introduce a novel deep learning component called variformer and integrate multi-omics data to capture intrinsic genomic features at both RNA and DNA levels. Results : Extensive experiments demonstrate that DeepRCI can detect RNA-chromatin interactions more accurately when compared to the state-of-the-art baseline prediction methods. Furthermore, the sequence features extracted by DeepRCI can be well matched to known critical gene regulatory components, indicating that our model can provide useful biological insights into understanding the underlying mechanisms of RNA-chromatin interactions. In addition, based on the prediction results, we further delineate the relationships between RNA-chromatin interactions and cellular functions, including gene expression and the modulation of cell states. Conclusions : In summary, DeepRCI can serve as a useful tool for characterizing RNA-chromatin interactions and studying the underlying gene regulatory code.
{"title":"DeepRCI: predicting RNA-chromatin interactions via deep learning with multi-omics data","authors":"Yuanpeng Xiong, Xuan He, Dan Zhao, Tao Jiang, Jianyang Zeng","doi":"10.15302/j-qb-022-0316","DOIUrl":"https://doi.org/10.15302/j-qb-022-0316","url":null,"abstract":"Background : Chromatin-associated RNA (caRNA) acts as a ubiquitous epigenetic layer in eukaryotes, and has been reported to be essential in various biological processes, including gene transcription, chromatin remodeling and cellular differentiation. Recently, numerous experimental techniques have been developed to characterize genome-wide RNA-chromatin interactions to understand their underlying biological functions. However, these experimental methods are generally expensive, time-consuming, and limited in identifying all potential sites, while most of the existing computational methods are restricted to detecting only specific types of RNAs interacting with chromatin. Methods : Here, we propose a highly interpretable computational framework, named DeepRCI, to identify the interactions between various types of RNAs and chromatin. In this framework, we introduce a novel deep learning component called variformer and integrate multi-omics data to capture intrinsic genomic features at both RNA and DNA levels. Results : Extensive experiments demonstrate that DeepRCI can detect RNA-chromatin interactions more accurately when compared to the state-of-the-art baseline prediction methods. Furthermore, the sequence features extracted by DeepRCI can be well matched to known critical gene regulatory components, indicating that our model can provide useful biological insights into understanding the underlying mechanisms of RNA-chromatin interactions. In addition, based on the prediction results, we further delineate the relationships between RNA-chromatin interactions and cellular functions, including gene expression and the modulation of cell states. Conclusions : In summary, DeepRCI can serve as a useful tool for characterizing RNA-chromatin interactions and studying the underlying gene regulatory code.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"1 1","pages":""},"PeriodicalIF":3.1,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67351638","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xiuquan Wang, Mian Umair Ahsan, Yunyun Zhou, Kai Wang
Background : Oxford Nanopore long-read sequencing technology addresses current limitations for DNA methylation detection that are inherent in short-read bisulfite sequencing or methylation microarrays. A number of analytical tools, such as Nanopolish, Guppy/Tombo and DeepMod, have been developed to detect DNA methylation on Nanopore data. However, additional improvements can be made in computational efficiency, prediction accuracy, and contextual interpretation on complex genomics regions (such as repetitive regions, low GC density regions). Method : In the current study, we apply Transformer architecture to detect DNA methylation on ionic signals from Oxford Nanopore sequencing data. Transformer is an algorithm that adopts self-attention architecture in the neural networks and has been widely used in natural language processing. Results : Compared to traditional deep-learning method such as convolutional neural network (CNN) and recurrent neural network (RNN), Transformer may have specific advantages in DNA methylation detection, because the self-attention mechanism can assist the relationship detection between bases that are far from each other and pay more attention to important bases that carry characteristic methylation-specific signals within a specific sequence context. Conclusion : We demonstrated the ability of Transformers to detect methylation on ionic signal data.
{"title":"Transformer-based DNA methylation detection on ionic signals from Oxford Nanopore sequencing data","authors":"Xiuquan Wang, Mian Umair Ahsan, Yunyun Zhou, Kai Wang","doi":"10.15302/j-qb-022-0323","DOIUrl":"https://doi.org/10.15302/j-qb-022-0323","url":null,"abstract":"Background : Oxford Nanopore long-read sequencing technology addresses current limitations for DNA methylation detection that are inherent in short-read bisulfite sequencing or methylation microarrays. A number of analytical tools, such as Nanopolish, Guppy/Tombo and DeepMod, have been developed to detect DNA methylation on Nanopore data. However, additional improvements can be made in computational efficiency, prediction accuracy, and contextual interpretation on complex genomics regions (such as repetitive regions, low GC density regions). Method : In the current study, we apply Transformer architecture to detect DNA methylation on ionic signals from Oxford Nanopore sequencing data. Transformer is an algorithm that adopts self-attention architecture in the neural networks and has been widely used in natural language processing. Results : Compared to traditional deep-learning method such as convolutional neural network (CNN) and recurrent neural network (RNN), Transformer may have specific advantages in DNA methylation detection, because the self-attention mechanism can assist the relationship detection between bases that are far from each other and pay more attention to important bases that carry characteristic methylation-specific signals within a specific sequence context. Conclusion : We demonstrated the ability of Transformers to detect methylation on ionic signal data.","PeriodicalId":45660,"journal":{"name":"Quantitative Biology","volume":"1 1","pages":""},"PeriodicalIF":3.1,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67351431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"生物学","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：