Haotian Teng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Haotian Teng has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Lachlan J. M. Coin
Department of Microbiology and Immunology, The University of Melbourne, at the Peter Doherty Institute for Infection and Immunity
6 shared publications - Ziv Bar-Joseph
Machine Learning Department, School of Computer Science, Carnegie Mellon University, Pittsburgh, USA
4 shared publications - Carl Kingsford
Ray and Stephanie Lane Computational Biology Department, School of Computer Science, Carnegie Mellon University, Pittsburgh, PA, USA
3 shared publications Benjamin Reames
University of Melbourne at The Peter Doherty Institute for Infection and Immunity Department of Microbiology and Immunology Melbourne, VIC, 3000
2 shared publications- Jessie J -Y Chang
Department of Microbiology and Immunology, University of Melbourne at the Peter Doherty Institute for Infection and Immunity
2 shared publications - Michael B. Hall
The University of Queensland
2 shared publications
Latest Publications
- Open Access
DTMol: Pocket-based Molecular Docking using Diffusion Transformers
In computational chemistry, molecular docking—predicting the binding structure of a small molecule ligand to a protein—is vital for understanding interactions between small molecules and their protein targets, with broad applications in drug discovery (Morris and Lim-Wilby, 2008). Traditional docking methods rely on energy-based scoring functions and optimization algorithms to identify ligand-binding structures. However, these methods are often slow and inaccurate due to the extensive search space for binding structures and the complexity of scoring function landscapes (Corso et al., 2023). Deep learning techniques have emerged as promising alternatives for molecular docking. These methods harness the power of neural networks to understand the complex interactions between protein pockets and ligands from large datasets. They can be classified into two categories: regression-based methods (Stark et al., 2022; Lu et al., 2022), which offer greater computational efficiency compared to traditional docking methods but have yet to offer substantial improvements in accuracy (Corso et al., 2023); and generative models, particularly diffusion models (Corso et al., 2023; Qiao et al., 2024; Nakata et al., 2023; Lu et al., 2024; Schneuing et al., 2024), which lead to significant accuracy improvements compared to regression-based methods. Most diffusion-based docking methods aim to address the blind docking problem, where the ligand is docked without prior knowledge of the specific protein pocket—a site of the protein with potential ligand-binding capabilities—and only the overall protein structure is provided (Yim et al., 2024). However, in many drug discovery cases, specific pockets have already been identified (Zheng et al., 2013), meaning that the docking process only needs to determine how the ligand binds to the pocket. This defines the pocket-based docking problem. Pocket-based docking leverages the physical and chemical information of protein side chains around the pocket, significantly reducing the search space for ligand binding structures and improving the understanding of atomic-level interactions. Additionally, focusing on the local protein substructure rather than the global structure enhances computational efficiency, which allows for the use of more complex neural architectures. We introduce DTMol, a novel diffusion model designed to tackle the pocket-based docking problem. Our model integrates a pretrained molecular representation framework with a diffusion transformer architecture. The advantages of this design are twofold: first, the molecular representation models we employ are pretrained on separate datasets of small molecules and protein pockets, which are larger than the datasets of ligand-pocket interactions. Consequently, these pretrained models encode typical structural features of both elements, enhancing our model’s ability to represent them and predict binding structures. Second, the transformer architecture used in our diffusion model is more capable of effectively capturing all-atom interactions between the pocket and the ligand. We test our method on the docking benchmark PDBBind v2020 and PoseBuster, and compare it with other methods. DTMol achieves 77.65% top-1 success rate (RMSD ¡ 2Å) on PoseBuster, outperforms the second-best docking software Gnina (65.65%) We also test the performance of our method on a real-world virtual screening task using Janus kinase 2 (PDB ID: 6BBV) as the target protein. A TR-FRET screening experiment is performed on the union of top ligands identified via physics-based docking and the diffusion model, in order to measure their inhibitory activity against the JAK2 protein. The results shows that our model achieved the only positive rank correlation score compared to other traditional docking methods and machine learning-based methods. To summarize, the main contributions of this work are: - We introduce DTMol, a novel diffusion model designed to address the pocket-based docking problem. This model represents the first application of the diffusion transformer architecture in solving pocket-based molecular docking. - We propose a novel SE(3)-equivariant transformer architecture with an SE(3)-invariant and a parallel SE(3)-equivariant flow, which is straightforward to implement and train while achieving excellent performance. - We achieve new state-of-the-art results on both the PDBBind and PoseBuster pocket-based docking benchmarks, obtaining top-1 prediction rates of 45.05% and 77.65%, respectively, at RMSD < 2Å. - Doing virtual screening on Janus kinase 2, DTMol is able to achieve the only positive rank score; and identifies 2 new experimentally effective JAK2 inhibitors.
Open AccessDTMol: Pocket-based Molecular Docking using Diffusion Transformers
In computational chemistry, molecular docking—predicting the binding structure of a small molecule ligand to a protein—is vital for understanding interactions between small molecules and their protein targets, with broad applications in drug discovery (Morris and Lim-Wilby, 2008). Traditional docking methods rely on energy-based scoring functions and optimization algorithms to identify ligand-binding structures. However, these methods are often slow and inaccurate due to the extensive search space for binding structures and the complexity of scoring function landscapes (Corso et al., 2023). Deep learning techniques have emerged as promising alternatives for molecular docking. These methods harness the power of neural networks to understand the complex interactions between protein pockets and ligands from large datasets. They can be classified into two categories: regression-based methods (Stark et al., 2022; Lu et al., 2022), which offer greater computational efficiency compared to traditional docking methods but have yet to offer substantial improvements in accuracy (Corso et al., 2023); and generative models, particularly diffusion models (Corso et al., 2023; Qiao et al., 2024; Nakata et al., 2023; Lu et al., 2024; Schneuing et al., 2024), which lead to significant accuracy improvements compared to regression-based methods. Most diffusion-based docking methods aim to address the blind docking problem, where the ligand is docked without prior knowledge of the specific protein pocket—a site of the protein with potential ligand-binding capabilities—and only the overall protein structure is provided (Yim et al., 2024). However, in many drug discovery cases, specific pockets have already been identified (Zheng et al., 2013), meaning that the docking process only needs to determine how the ligand binds to the pocket. This defines the pocket-based docking problem. Pocket-based docking leverages the physical and chemical information of protein side chains around the pocket, significantly reducing the search space for ligand binding structures and improving the understanding of atomic-level interactions. Additionally, focusing on the local protein substructure rather than the global structure enhances computational efficiency, which allows for the use of more complex neural architectures. We introduce DTMol, a novel diffusion model designed to tackle the pocket-based docking problem. Our model integrates a pretrained molecular representation framework with a diffusion transformer architecture. The advantages of this design are twofold: first, the molecular representation models we employ are pretrained on separate datasets of small molecules and protein pockets, which are larger than the datasets of ligand-pocket interactions. Consequently, these pretrained models encode typical structural features of both elements, enhancing our model’s ability to represent them and predict binding structures. Second, the transformer architecture used in our diffusion model is more capable of effectively capturing all-atom interactions between the pocket and the ligand. We test our method on the docking benchmark PDBBind v2020 and PoseBuster, and compare it with other methods. DTMol achieves 77.65% top-1 success rate (RMSD ¡ 2Å) on PoseBuster, outperforms the second-best docking software Gnina (65.65%) We also test the performance of our method on a real-world virtual screening task using Janus kinase 2 (PDB ID: 6BBV) as the target protein. A TR-FRET screening experiment is performed on the union of top ligands identified via physics-based docking and the diffusion model, in order to measure their inhibitory activity against the JAK2 protein. The results shows that our model achieved the only positive rank correlation score compared to other traditional docking methods and machine learning-based methods. To summarize, the main contributions of this work are: - We introduce DTMol, a novel diffusion model designed to address the pocket-based docking problem. This model represents the first application of the diffusion transformer architecture in solving pocket-based molecular docking. - We propose a novel SE(3)-equivariant transformer architecture with an SE(3)-invariant and a parallel SE(3)-equivariant flow, which is straightforward to implement and train while achieving excellent performance. - We achieve new state-of-the-art results on both the PDBBind and PoseBuster pocket-based docking benchmarks, obtaining top-1 prediction rates of 45.05% and 77.65%, respectively, at RMSD < 2Å. - Doing virtual screening on Janus kinase 2, DTMol is able to achieve the only positive rank score; and identifies 2 new experimentally effective JAK2 inhibitors.
- Open Access
- 6 January 2025
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 14
Using synthetic RNA to benchmark poly(A) length inference from direct RNA sequencing
Polyadenylation is a dynamic process that is important in cellular physiology, which has implications in messenger RNA decay rates, translation efficiency, and isoform-specific regulation. Oxford Nanopore Technologies direct RNA sequencing provides a strategy for sequencing the full-length RNA molecule and analysis of the transcriptome. Several tools are currently available for poly(A) tail length estimation, including well-established methods like tailfindr and nanopolish, as well as more recent deep learning models like Dorado. However, there has been limited benchmarking of the accuracy of these tools against gold-standard datasets. In this article, we present our novel deep learning poly(A) estimation tool—BoostNano—and compare with 3 existing tools—tailfindr, nanopolish, and Dorado. We evaluate the 4 poly(A) estimation tools, using 2 sets of synthetic in vitro transcribed RNA standards with known poly(A) tail lengths—Sequin (30 or 60 nucleotides) and enhanced green fluorescent protein (10–150 nucleotides) RNA. Analyzing datasets with known ground-truth values is a valuable approach to measuring the accuracy of poly(A) length estimation. The tools demonstrated length- and sample-dependent performance, and accuracy was enhanced by averaging over multiple reads via estimation of the peak of the density distribution. Overall, Dorado is recommended as the preferred approach due to its relatively fast runtimes, low mean error, and ease of use with integration with base-calling. These results provide a reference for poly(A) tail length estimation analysis, aiding in improving our understanding of the transcriptome and the relationship between poly(A) tail length and other transcriptional mechanisms, including transcript stability or quantification.
- 6 January 2025
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 14
Open AccessUsing synthetic RNA to benchmark poly(A) length inference from direct RNA sequencing
Polyadenylation is a dynamic process that is important in cellular physiology, which has implications in messenger RNA decay rates, translation efficiency, and isoform-specific regulation. Oxford Nanopore Technologies direct RNA sequencing provides a strategy for sequencing the full-length RNA molecule and analysis of the transcriptome. Several tools are currently available for poly(A) tail length estimation, including well-established methods like tailfindr and nanopolish, as well as more recent deep learning models like Dorado. However, there has been limited benchmarking of the accuracy of these tools against gold-standard datasets. In this article, we present our novel deep learning poly(A) estimation tool—BoostNano—and compare with 3 existing tools—tailfindr, nanopolish, and Dorado. We evaluate the 4 poly(A) estimation tools, using 2 sets of synthetic in vitro transcribed RNA standards with known poly(A) tail lengths—Sequin (30 or 60 nucleotides) and enhanced green fluorescent protein (10–150 nucleotides) RNA. Analyzing datasets with known ground-truth values is a valuable approach to measuring the accuracy of poly(A) length estimation. The tools demonstrated length- and sample-dependent performance, and accuracy was enhanced by averaging over multiple reads via estimation of the peak of the density distribution. Overall, Dorado is recommended as the preferred approach due to its relatively fast runtimes, low mean error, and ease of use with integration with base-calling. These results provide a reference for poly(A) tail length estimation analysis, aiding in improving our understanding of the transcriptome and the relationship between poly(A) tail length and other transcriptional mechanisms, including transcript stability or quantification.
- Open Access
Using synthetic RNA to benchmark poly(A) length inference from direct RNA sequencing
Polyadenylation is a dynamic process which is important in cellular physiology. Oxford Nanopore Technologies direct RNA-sequencing provides a strategy for sequencing the full-length RNA molecule and analysis of the transcriptome and epi-transcriptome. There are currently several tools available for poly(A) tail-length estimation, including well-established tools such as tailfindr and nanopolish , as well as two more recent deep learning models: Dorado and BoostNano . However, there has been limited benchmarking of the accuracy of these tools against gold-standard datasets. In this paper we evaluate four poly(A) estimation tools using synthetic RNA standards (Sequins), which have known poly(A) tail-lengths and provide a valuable approach to measuring the accuracy of poly(A) tail-length estimation. All four tools generate mean tail-length estimates which lie within 12% of the correct value. Overall, Dorado is recommended as the preferred approach due to its relatively fast run times, low coefficient of variation and ease of use with integration with base-calling.
Open AccessUsing synthetic RNA to benchmark poly(A) length inference from direct RNA sequencing
Polyadenylation is a dynamic process which is important in cellular physiology. Oxford Nanopore Technologies direct RNA-sequencing provides a strategy for sequencing the full-length RNA molecule and analysis of the transcriptome and epi-transcriptome. There are currently several tools available for poly(A) tail-length estimation, including well-established tools such as tailfindr and nanopolish , as well as two more recent deep learning models: Dorado and BoostNano . However, there has been limited benchmarking of the accuracy of these tools against gold-standard datasets. In this paper we evaluate four poly(A) estimation tools using synthetic RNA standards (Sequins), which have known poly(A) tail-lengths and provide a valuable approach to measuring the accuracy of poly(A) tail-length estimation. All four tools generate mean tail-length estimates which lie within 12% of the correct value. Overall, Dorado is recommended as the preferred approach due to its relatively fast run times, low coefficient of variation and ease of use with integration with base-calling.
- Open Access
- 15 October 2024
- Published by Cold Spring Harbor Laboratory in Genome Research
- Vol. 34 (11) , 1987-1999
Detecting m6A RNA modification from nanopore sequencing using a semisupervised learning framework
Direct nanopore-based RNA sequencing can be used to detect posttranscriptional base modifications, such as N6-methyladenosine (m6A) methylation, based on the electric current signals produced by the distinct chemical structures of modified bases. A key challenge is the scarcity of adequate training data with known methylation modifications. We present Xron, a hybrid encoder–decoder framework that delivers a direct methylation-distinguishing basecaller by training on synthetic RNA data and immunoprecipitation (IP)-based experimental data in two steps. First, we generate data with more diverse modification combinations through in silico cross-linking. Second, we use this data set to train an end-to-end neural network basecaller followed by fine-tuning on IP-based experimental data with label smoothing. The trained neural network basecaller outperforms existing methylation detection methods on both read-level and site-level prediction scores. Xron is a standalone, end-to-end m6A-distinguishing basecaller capable of detecting methylated bases directly from raw sequencing signals, enabling de novo methylome assembly.
- 15 October 2024
- Published by Cold Spring Harbor Laboratory in Genome Research
- Vol. 34 (11) , 1987-1999
Open AccessDetecting m6A RNA modification from nanopore sequencing using a semisupervised learning framework
Direct nanopore-based RNA sequencing can be used to detect posttranscriptional base modifications, such as N6-methyladenosine (m6A) methylation, based on the electric current signals produced by the distinct chemical structures of modified bases. A key challenge is the scarcity of adequate training data with known methylation modifications. We present Xron, a hybrid encoder–decoder framework that delivers a direct methylation-distinguishing basecaller by training on synthetic RNA data and immunoprecipitation (IP)-based experimental data in two steps. First, we generate data with more diverse modification combinations through in silico cross-linking. Second, we use this data set to train an end-to-end neural network basecaller followed by fine-tuning on IP-based experimental data with label smoothing. The trained neural network basecaller outperforms existing methylation detection methods on both read-level and site-level prediction scores. Xron is a standalone, end-to-end m6A-distinguishing basecaller capable of detecting methylated bases directly from raw sequencing signals, enabling de novo methylome assembly.
- Open Access
Detecting m6A RNA modification from nanopore sequencing using a semi-supervised learning framework
Direct nanopore-based RNA sequencing can be used to detect post-transcriptional base modifications, such as m6A methylation, based on the electric current signals produced by the distinct chemical structures of modified bases. A key challenge is the scarcity of adequate training data with known methylation modifications. We present Xron, a hybrid encoder-decoder framework that delivers a direct methylation-distinguishing basecaller by training on synthetic RNA data and immunoprecipitation-based experimental data in two steps. First, we generate data with more diverse modification combinations through in silico cross-linking. Second, we use this dataset to train an end-to-end neural network basecaller followed by fine-tuning on immunoprecipitation-based experimental data with label-smoothing. The trained neural network basecaller outperforms existing methylation detection methods on both read-level and site-level prediction scores. Xron is a standalone, end-to-end m6A-distinguishing basecaller capable of detecting methylated bases directly from raw sequencing signals, enabling de novo methylome assembly.
Open AccessDetecting m6A RNA modification from nanopore sequencing using a semi-supervised learning framework
Direct nanopore-based RNA sequencing can be used to detect post-transcriptional base modifications, such as m6A methylation, based on the electric current signals produced by the distinct chemical structures of modified bases. A key challenge is the scarcity of adequate training data with known methylation modifications. We present Xron, a hybrid encoder-decoder framework that delivers a direct methylation-distinguishing basecaller by training on synthetic RNA data and immunoprecipitation-based experimental data in two steps. First, we generate data with more diverse modification combinations through in silico cross-linking. Second, we use this dataset to train an end-to-end neural network basecaller followed by fine-tuning on immunoprecipitation-based experimental data with label-smoothing. The trained neural network basecaller outperforms existing methylation detection methods on both read-level and site-level prediction scores. Xron is a standalone, end-to-end m6A-distinguishing basecaller capable of detecting methylated bases directly from raw sequencing signals, enabling de novo methylome assembly.
- Open Access
- 8 October 2021
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 38 (4) , 997-1004
Clustering spatial transcriptomics data
Recent advancements in fluorescence in situ hybridization (FISH) techniques enable them to concurrently obtain information on the location and gene expression of single cells. A key question in the initial analysis of such spatial transcriptomics data is the assignment of cell types. To date, most studies used methods that only rely on the expression levels of the genes in each cell for such assignments. To fully utilize the data and to improve the ability to identify novel sub-types, we developed a new method, FICT, which combines both expression and neighborhood information when assigning cell types. FICT optimizes a probabilistic function that we formalize and for which we provide learning and inference algorithms. We used FICT to analyze both simulated and several real spatial transcriptomics data. As we show, FICT can accurately identify cell types and sub-types, improving on expression only methods and other methods proposed for clustering spatial transcriptomics data. Some of the spatial sub-types identified by FICT provide novel hypotheses about the new functions for excitatory and inhibitory neurons. FICT is available at: https://github.com/haotianteng/FICT. Supplementary data are available at Bioinformatics online.
- 8 October 2021
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 38 (4) , 997-1004
Open AccessClustering spatial transcriptomics data
Recent advancements in fluorescence in situ hybridization (FISH) techniques enable them to concurrently obtain information on the location and gene expression of single cells. A key question in the initial analysis of such spatial transcriptomics data is the assignment of cell types. To date, most studies used methods that only rely on the expression levels of the genes in each cell for such assignments. To fully utilize the data and to improve the ability to identify novel sub-types, we developed a new method, FICT, which combines both expression and neighborhood information when assigning cell types. FICT optimizes a probabilistic function that we formalize and for which we provide learning and inference algorithms. We used FICT to analyze both simulated and several real spatial transcriptomics data. As we show, FICT can accurately identify cell types and sub-types, improving on expression only methods and other methods proposed for clustering spatial transcriptomics data. Some of the spatial sub-types identified by FICT provide novel hypotheses about the new functions for excitatory and inhibitory neurons. FICT is available at: https://github.com/haotianteng/FICT. Supplementary data are available at Bioinformatics online.
- Open Access
Cell Type Assignments for Spatial Transcriptomics Data
Motivation: Recent advancements in fluorescence in situ hybridization (FISH) techniques enable them to concurrently obtain information on the location and gene expression of single cells. A key question in the initial analysis of such spatial transcriptomics data is the assignment of cell types. To date, most studies used methods that only rely on the expression levels of the genes in each cell for such assignments. To fully utilize the data and to improve the ability to identify novel sub-types we developed a new method, FICT, which combines both expression and neighborhood information when assigning cell types. Results: FICT optimizes a probabilistic function that we formalize and for which we provide learning and inference algorithms. We used FICT to analyze both simulated and several real spatial transcriptomics data. As we show, FICT can accurately identify cell types and sub-types improving on expression only methods and other methods proposed for clustering spatial transcriptomics data. Some of the spatial sub-types identified by FICT provide novel hypotheses about the new functions for excitatory and inhibitory neurons. Availability: FICT is available at: https://github.com/haotianteng/FICT Contact: zivbj@andrew.cmu.edu
Open AccessCell Type Assignments for Spatial Transcriptomics Data
Motivation: Recent advancements in fluorescence in situ hybridization (FISH) techniques enable them to concurrently obtain information on the location and gene expression of single cells. A key question in the initial analysis of such spatial transcriptomics data is the assignment of cell types. To date, most studies used methods that only rely on the expression levels of the genes in each cell for such assignments. To fully utilize the data and to improve the ability to identify novel sub-types we developed a new method, FICT, which combines both expression and neighborhood information when assigning cell types. Results: FICT optimizes a probabilistic function that we formalize and for which we provide learning and inference algorithms. We used FICT to analyze both simulated and several real spatial transcriptomics data. As we show, FICT can accurately identify cell types and sub-types improving on expression only methods and other methods proposed for clustering spatial transcriptomics data. Some of the spatial sub-types identified by FICT provide novel hypotheses about the new functions for excitatory and inhibitory neurons. Availability: FICT is available at: https://github.com/haotianteng/FICT Contact: zivbj@andrew.cmu.edu
- Open Access
- 1 February 2020
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 9 (2)
Evaluating the genome and resistome of extensively drug-resistant Klebsiella pneumoniae using native DNA and RNA Nanopore sequencing
Klebsiella pneumoniae frequently harbours multidrug resistance, and current diagnostics struggle to rapidly identify appropriate antibiotics to treat these bacterial infections. The MinION device can sequence native DNA and RNA in real time, providing an opportunity to compare the utility of DNA and RNA for prediction of antibiotic susceptibility. However, the effectiveness of bacterial direct RNA sequencing and base-calling has not previously been investigated. This study interrogated the genome and transcriptome of 4 extensively drug-resistant (XDR) K. pneumoniae clinical isolates; however, further antimicrobial susceptibility testing identified 3 isolates as pandrug-resistant (PDR). The majority of acquired resistance (≥75%) resided on plasmids including several megaplasmids (≥100 kb). DNA sequencing detected most resistance genes (≥70%) within 2 hours of sequencing. Neural network–based base-calling of direct RNA achieved up to 86% identity rate, although ≤23% of reads could be aligned. Direct RNA sequencing (with ∼6 times slower pore translocation) was able to identify (within 10 hours) ≥35% of resistance genes, including those associated with resistance to aminoglycosides, β-lactams, trimethoprim, and sulphonamide and also quinolones, rifampicin, fosfomycin, and phenicol in some isolates. Direct RNA sequencing also identified the presence of operons containing up to 3 resistance genes. Polymyxin-resistant isolates showed a heightened transcription of phoPQ (≥2-fold) and the pmrHFIJKLM operon (≥8-fold). Expression levels estimated from direct RNA sequencing displayed strong correlation (Pearson: 0.86) compared to quantitative real-time PCR across 11 resistance genes. Overall, MinION sequencing rapidly detected the XDR/PDR K. pneumoniae resistome, and direct RNA sequencing provided accurate estimation of expression levels of these genes.
- 1 February 2020
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 9 (2)
Open AccessEvaluating the genome and resistome of extensively drug-resistant Klebsiella pneumoniae using native DNA and RNA Nanopore sequencing
Klebsiella pneumoniae frequently harbours multidrug resistance, and current diagnostics struggle to rapidly identify appropriate antibiotics to treat these bacterial infections. The MinION device can sequence native DNA and RNA in real time, providing an opportunity to compare the utility of DNA and RNA for prediction of antibiotic susceptibility. However, the effectiveness of bacterial direct RNA sequencing and base-calling has not previously been investigated. This study interrogated the genome and transcriptome of 4 extensively drug-resistant (XDR) K. pneumoniae clinical isolates; however, further antimicrobial susceptibility testing identified 3 isolates as pandrug-resistant (PDR). The majority of acquired resistance (≥75%) resided on plasmids including several megaplasmids (≥100 kb). DNA sequencing detected most resistance genes (≥70%) within 2 hours of sequencing. Neural network–based base-calling of direct RNA achieved up to 86% identity rate, although ≤23% of reads could be aligned. Direct RNA sequencing (with ∼6 times slower pore translocation) was able to identify (within 10 hours) ≥35% of resistance genes, including those associated with resistance to aminoglycosides, β-lactams, trimethoprim, and sulphonamide and also quinolones, rifampicin, fosfomycin, and phenicol in some isolates. Direct RNA sequencing also identified the presence of operons containing up to 3 resistance genes. Polymyxin-resistant isolates showed a heightened transcription of phoPQ (≥2-fold) and the pmrHFIJKLM operon (≥8-fold). Expression levels estimated from direct RNA sequencing displayed strong correlation (Pearson: 0.86) compared to quantitative real-time PCR across 11 resistance genes. Overall, MinION sequencing rapidly detected the XDR/PDR K. pneumoniae resistome, and direct RNA sequencing provided accurate estimation of expression levels of these genes.
- Open Access
- 1 May 2019
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 8 (5)
Correction to: Chiron: translating nanopore raw signal directly into nucleotide sequence using deep learning
- 1 May 2019
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 8 (5)
Open AccessCorrection to: Chiron: translating nanopore raw signal directly into nucleotide sequence using deep learning
- Open Access
- 10 April 2018
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 7 (5)
Chiron: translating nanopore raw signal directly into nucleotide sequence using deep learning
Sequencing by translocating DNA fragments through an array of nanopores is a rapidly maturing technology that offers faster and cheaper sequencing than other approaches. However, accurately deciphering the DNA sequence from the noisy and complex electrical signal is challenging. Here, we report Chiron, the first deep learning model to achieve end-to-end basecalling and directly translate the raw signal to DNA sequence without the error-prone segmentation step. Trained with only a small set of 4,000 reads, we show that our model provides state-of-the-art basecalling accuracy, even on previously unseen species. Chiron achieves basecalling speeds of more than 2,000 bases per second using desktop computer graphics processing units.
- 10 April 2018
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 7 (5)
Open AccessChiron: translating nanopore raw signal directly into nucleotide sequence using deep learning
Sequencing by translocating DNA fragments through an array of nanopores is a rapidly maturing technology that offers faster and cheaper sequencing than other approaches. However, accurately deciphering the DNA sequence from the noisy and complex electrical signal is challenging. Here, we report Chiron, the first deep learning model to achieve end-to-end basecalling and directly translate the raw signal to DNA sequence without the error-prone segmentation step. Trained with only a small set of 4,000 reads, we show that our model provides state-of-the-art basecalling accuracy, even on previously unseen species. Chiron achieves basecalling speeds of more than 2,000 bases per second using desktop computer graphics processing units.