IP Library › Granted Patent US 12,437,184
Granted Patent B2
US 12,437,184 · App. 16/610,897 · Granted Oct 7, 2025

Machine learning analysis of nanopore measurements

Inventors: Timothy Lee Massingham (Oxford, GB); Joseph Edward Harvey (Oxford, GB)
Assignee: Oxford Nanopore Technologies PLC
G06N3/044G01N33/48721G06F18/22G06N7/01C12Q1/6869
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,437,184
App. No.
16/610,897
Granted
Oct 7, 2025
Kind
B2
Abstract

A series of measurements taken from a polymer during translocation through a nanopore is analysed using a machine learning technique using a recurrent neural network (RNN). The RNN may derive posterior probability matrices each representing, in respect of different respective historical sequences of polymer units corresponding to measurements prior to the respective measurement, posterior probabilities of plural different changes to the respective historical sequence of polymer units giving rise to a new sequence of polymer units. Alternatively, the RNN may output decisions on the identity of successive polymer units of the series of polymer units, wherein the decisions are fed back into the recurrent neural network. The analysis may comprise performing convolutions of groups of consecutive measurements using a trained feature detector such as a convolutional neural network to derive a series of feature vectors, on which the RNN operates.

Claims (44)

1. A method of high-rate sequencing of polymers using a nanopore measurement and analysis system, the method comprising:

placing a polymer into the nanopore measurement and analysis system; and

sequencing the polymer using the nanopore measurement and analysis system at least in part by:

translocating at least a portion of the polymer through a nanopore of the nanopore measurement and analysis system at a sequencing rate in the range of 10-1000 polymer units per second, wherein the sequencing rate reflects a rate at which the polymer translocates through the nanopore;

measuring, using the nanopore measurement and analysis system, electrical signals generated during the translocating of the polymer through the nanopore, at a sampling rate greater than or equal to the sequencing rate, to generate a series of measurements;

estimating, using a convolutional neural network comprising a convolutional layer and a recurrent neural network comprising a bidirectional recurrent layer, the bidirectional recurrent layer comprising a plurality of long short-term memory (LSTM) units, a series of polymer units within the polymer at least in part by:

converting the series of measurements into a series of feature vectors by applying the convolutional neural network comprising the convolutional layer to a series of groups of measurements derived from the series of measurements, wherein the series of feature vectors include a first feature vector corresponding to a first group of measurements in the series of groups of measurements;

processing the series of feature vectors using the recurrent neural network comprising the bidirectional recurrent layer to obtain respective outputs including a first output corresponding to a first feature vector in the series of feature vectors, wherein the first output represents, in respect of different respective historical sequences of polymer units corresponding to measurements obtained prior or subsequent to the first group of measurements, posterior probabilities of plural different changes to the respective historical sequences of polymer units; and

generating the estimate of the series of polymer units within the polymer using the respective outputs obtained using the recurrent neural network comprising the bidirectional recurrent layer.

2. The method according to claim 1 , wherein generating the estimate of the series of polymer units using the respective outputs is performed by estimating likelihoods of paths through the respective outputs and identifying a path based on the likelihoods.

3. The method according to claim 1 , wherein generating the estimate of the series of polymer units is performed by selecting one of a set of plural reference series of polymer units to which the series of polymer units of the polymer are most similar.

4. The method according to claim 1 , wherein generating the estimate of the series of polymer units is performed by estimating differences between the series of polymer units of the polymer and a reference series of polymer units from the respective outputs.

5. The method according to claim 1 , wherein the estimate is an estimate of whether part of the series of polymer units of the polymer is a reference series of polymer units.

6. The method according to claim 1 , further comprising deriving a score in respect of at least one reference series of polymer units representing a probability of the series of polymer units of the polymer being the reference series of polymer units.

7. The method according to claim 1 , wherein the plural different changes include changes that remove a single polymer unit from a beginning or end of a respective historical sequence of polymer units and add a single polymer unit to the end or beginning of the respective historical sequence of polymer units.

8. The method according to claim 1 , wherein the plural different changes include changes that remove two or more polymer units from beginning or end of a respective historical sequence of polymer units and add two or more polymer units to the end or beginning of the respective historical sequence of polymer units.

9. The method according to claim 1 , wherein the groups of measurements are overlapping groups of measurements.

10. The method of claim 1 , wherein the convolutional neural network further comprises a pooling layer, and wherein outputs of the convolutional layer are inputs into the pooling layer.

11. The method of claim 1 , wherein the bidirectional recurrent layer comprises a first unidirectional recurrent layer comprising multiple LSTM units of the plurality of LSTM units connected in a first direction, and a second unidirectional recurrent layer comprising multiple LSTM units of the plurality of LSTM units connected in a second direction, opposite the first direction, and wherein outputs of the first unidirectional recurrent layer are inputs into the second unidirectional recurrent layer.

12. The method of claim 11 , wherein processing the series of feature vectors using the recurrent neural network comprises updating state vectors of LSTM units of the first unidirectional recurrent layer based on the feature vectors and state vectors of LSTM units preceding the multiple LSTM units in the first unidirectional recurrent layer, and updating state vectors of LSTM units of the second unidirectional recurrent layer based on outputs of the first unidirectional recurrent layer and state vectors of LSTM units preceding the multiple LSTM units in the second unidirectional recurrent layer.

13. The method of claim 1 , wherein the polymer is a polynucleotide, and contains at least 30 kilobases (kB).

14. A nanopore measurement and analysis system, comprising:

a nanopore measurement system configured to perform:

sequencing a polymer using the nanopore measurement and analysis system at least in part by:

translocating at least a portion of the polymer through a nanopore of the nanopore measurement and analysis system at a sequencing rate in the range of 10-1000 polymer units per second, wherein the sequencing rate reflects a rate at which the polymer translocates through the nanopore;

measuring electrical signals generated during the translocating of the polymer through the nanopore, at a sampling rate greater than or equal to the sequencing rate, to generate a series of measurements; and

at least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by at least one processor of an analysis system, causes the at least processor to perform:

further sequencing the polymer, at least in part by:

estimating, using a convolutional neural network comprising a convolutional layer and a recurrent neural network comprising a bidirectional recurrent layer, the bidirectional recurrent layer comprising a plurality of long short-term memory (LSTM) units, a series of polymer units within the polymer at least in part by:

converting the series of measurements into a series of feature vectors by applying the convolutional neural network comprising the convolutional layer to a series of groups of measurements derived from the series of measurements, the series of feature vectors including a first feature vector corresponding to a first group of measurements in the series of groups of measurements;

processing the series of feature vectors using the recurrent neural network comprising the bidirectional recurrent layer to obtain respective outputs, including a first output corresponding to a first feature vector in the series of feature vectors, wherein the first output represents, in respect of different respective historical sequences of polymer units corresponding to measurements obtained prior or subsequent to the first group of measurements, posterior probabilities of plural different changes to the respective historical sequences of polymer units; and

generating the estimate of the series of polymer units within the polymer using the respective outputs obtained using the recurrent neural network comprising the bidirectional recurrent layer.

15. The system of claim 14 , wherein generating the estimate of the series of polymer units using the respective outputs is performed by estimating likelihoods of paths through the respective outputs and identifying a path based on the likelihoods.

16. The system of claim 14 , wherein the plural different changes include changes that remove a single polymer unit from a beginning or end of a historical sequence of polymer units and add a single polymer unit to the end or beginning of the respective historical sequence of polymer units.

17. The system of claim 14 , wherein the bidirectional recurrent layer comprises a first unidirectional recurrent layer comprising multiple LSTM units of the plurality of LSTM units connected in a first direction, and a second unidirectional recurrent layer comprising multiple LSTM units of the plurality of LSTM units connected in a second direction, opposite the first direction, and wherein outputs of the first unidirectional recurrent layer are inputs into the second unidirectional recurrent layer.

18. At least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by at least one computer hardware processor coupled to a nanopore measurement and analysis system, causes the at least one computer hardware processor and the nanopore measurement and analysis system to sequence a polymer at least in part by:

sequencing the polymer using the nanopore measurement and analysis system at least in part by:

obtaining a series of measurements generated from electrical signals measured during translocation of a polymer through a nanopore of a nanopore measurement system at a sequencing rate in the range of 10-1000 polymer units per second, wherein the sequencing rate reflects a rate at which the polymer translocates through the nanopore and wherein the series of measurements are obtained at a sampling rate greater than or equal to the sequencing rate;

estimating, using a convolutional neural network comprising a convolutional layer and a recurrent neural network comprising a bidirectional recurrent layer, the bidirectional recurrent layer comprising a plurality of long short-term memory (LSTM) units, a series of polymer units within the polymer at least in part by:

converting the series of measurements into a series of feature vectors by applying the convolutional neural network comprising the convolutional layer to a series of groups of measurements derived from the series of measurements, the series of feature vectors including a first feature vector corresponding to a first group of measurements in the series of groups of measurements;

processing the series of feature vectors using the recurrent neural network comprising the bidirectional recurrent layer to obtain respective outputs including a first output corresponding to a first feature vector in the series of feature vectors, wherein the first output represents, in respect of different respective historical sequences of polymer units corresponding to measurements obtained prior or subsequent to the first group of measurements, posterior probabilities of plural different changes to the respective historical sequences of polymer units; and

generating the estimate of the series of polymer units within the polymer using the respective outputs obtained using the recurrent neural network comprising the bidirectional recurrent layer.

19. The at least one non-transitory computer-readable storage medium of claim 18 , wherein the plural different changes include changes that remove a single polymer unit from a beginning or end of a respective historical sequence of polymer units and add a single polymer unit to the end or beginning of the respective historical sequence of polymer units.

20. The at least one non-transitory computer-readable medium of claim 18 , wherein the bidirectional recurrent layer comprises a first unidirectional recurrent layer comprising multiple LSTM units of the plurality of LSTM units connected in a first direction, and a second unidirectional recurrent layer comprising multiple LSTM units of the plurality of LSTM units connected in a second direction, opposite the first direction, and wherein outputs of the first unidirectional recurrent layer are inputs into the second unidirectional recurrent layer.

Assignments (2)
CHANGE OF NAME Recorded Jan 14, 2022
From: OXFORD NANOPORE TECHNOLOGIES LIMITED
To: OXFORD NANOPORE TECHNOLOGIES PLC
Reel/Frame 058737/0664 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 2, 2020
From: MASSINGHAM, TIMOTHY LEE; HARVEY, JOSEPH EDWARD
To: OXFORD NANOPORE TECHNOLOGIES LIMITED
Reel/Frame 052810/0117 →
Priority Claims (1)
GB 1707138 · May 4, 2017 · national
Continuity (2)
Related Publication 20200309761A1 · Oct 1, 2020
Related Publication 20240370696A2 · Nov 7, 2024
References Cited (98)
US 6087099A · Gupte et al. · 2000 [cited by applicant]
US 8486630B2 · Pan et al. · 2013 [cited by applicant]
US 10017814B2 · Gundlach et al. · 2018 [cited by applicant]
US 20120322692A1 · Pham et al. · 2012 [cited by applicant]
US 20130217006A1 · Sorenson et al. · 2013 [cited by applicant]
US 20130327644A1 · Turner et al. · 2013 [cited by applicant]
US 20140052381A1 · Utiramerur et al. · 2014 [cited by applicant]
US 20150057948A1 · Reid et al. · 2015 [cited by applicant]
US 20170067101A1 · Clarke et al. · 2017 [cited by applicant]
US 20180164280A1 · Ecker et al. · 2018 [cited by applicant]
US 20180179591A1 · Belgrader et al. · 2018 [cited by applicant]
US 20200176082A1 · Massingham · 2020 [cited by applicant]
US 20220213541A1 · Brown et al. · 2022 [cited by applicant]
AU 2010326349A1 · 2012 [cited by applicant]
CN 101680873A · 2010 [cited by applicant]
CN 104321441A · 2015 [cited by applicant]
WO WO2000028312A1 · 2000 [cited by applicant]
WO WO2000079257A1 · 2000 [cited by applicant]
WO WO2005124888A1 · 2005 [cited by applicant]
WO WO2008102120A2 · 2008 [cited by applicant]
WO WO2008102210A2 · 2008 [cited by applicant]
WO WO2008124107A1 · 2008 [cited by applicant]
WO WO2009035647A1 · 2009 [cited by applicant]
WO WO2009077734A2 · 2009 [cited by applicant]
WO WO2010086603A1 · 2010 [cited by applicant]
WO WO2010109197A2 · 2010 [cited by applicant]
WO WO2010122293A1 · 2010 [cited by applicant]
WO WO2011046706A1 · 2011 [cited by applicant]
WO WO2011067559A1 · 2011 [cited by applicant]
WO WO2012005857A1 · 2012 [cited by applicant]
WO WO2012033524A2 · 2012 [cited by applicant]
WO WO2012107778A2 · 2012 [cited by applicant]
WO WO2012138357A1 · 2012 [cited by applicant]
WO WO2013014451A1 · 2013 [cited by applicant]
WO WO2013041878A1 · 2013 [cited by applicant]
WO WO2013083983A1 · 2013 [cited by applicant]
WO WO2013121224A1 · 2013 [cited by applicant]
WO WO2013153359A1 · 2013 [cited by applicant]
WO WO2014064443A2 · 2014 [cited by applicant]
WO WO2014064444A1 · 2014 [cited by applicant]
WO WO2015124935A1 · 2015 [cited by applicant]
WO WO2015140535A1 · 2015 [cited by applicant]
WO WO2016034591A2 · 2016 [cited by applicant]
WO WO2016181118A1 · 2016 [cited by applicant]
WO WO2016187519A1 · 2016 [cited by applicant]
WO WO2018115855A1 · 2018 [cited by applicant]
WO WO2018203084A1 · 2018 [cited by applicant]
WO WO2019220094A1 · 2019 [cited by applicant]
Boža, Vladimír, Broňa Brejová, and TomášVinař. “DeepNano: Deep Recurrent Neural Networks for Base Calling in MinION Nanopore Reads.” arXiv preprint arXiv:1603.09195 (2016). (Year: 2016). [cited by examiner]
Rand, Arthur C., et al. “Mapping DNA methylation with high-throughput nanopore sequencing.” Nature methods 14.4 (2017): 411-413. (Year: 2017). [cited by examiner]
Rand Supplemental “Mapping DNA methylation with high-throughput nanopore sequencing.” Nature methods 14.4 (2017): 411-413. (Year: 2017). [cited by examiner]
De Mulder, Wim, Steven Bethard, and Marie-Francine Moens. “A survey on the application of recurrent neural networks to statistical language modeling.” Computer Speech & Language 30.1 (2015): 61-98. (Year: 2015). [cited by examiner]
Deamer, David, Mark Akeson, and Daniel Branton. “Three decades of nanopore sequencing.” Nature biotechnology 34.5 (2016):518-524. (Year: 2016). [cited by examiner]
International Search Report and Written Opinion from Application No. PCT/GB2018/051208, mailed Jul. 24, 2018. [cited by applicant]
International Preliminary Report on Patentability for Application No. PCT/GB2018/051208, mailed Nov. 14, 2019. [cited by applicant]
Amodei et al., Deep Speech 2: End-to-End Speech Recognition in English and Mandarin Baidu Research—Silicon Valley AI Lab. Dec. 8, 2015. Retrieved from https://arxiv.org/pdf/1512.02595v1.pdf. 28 pages. [cited by applicant]
Boža et al., DeepNano: Deep recurrent neural networks for base calling in MinION nanopore reads. Corr ARXIV. Mar. 30, 2016. 12 pages. Author Manuscript. [cited by applicant]
Boža et al., DeepNano: Deep recurrent neural networks for base calling in MinION nanopore reads. PLoS One. Jun. 5, 2017;12(6):e0178751. doi: 10.1371/journal.pone.0178751. eCollection 2017. [cited by applicant]
Butler et al., Single-molecule DNA detection with an engineered MspA protein nanopore. Proc Natl Acad Sci U S A. Dec. 30, 2008;105(52):20647-52. doi: 10.1073/pnas.0807514106. Epub Dec. 19, 2008. [cited by applicant]
Fernández et al., An Application of Recurrent Neural Networks to Discriminative Keyword Spotting. ICANN 2007: Artificial Neural Networks. pp. 220-229. 2007. 10 pages. [cited by applicant]
Gers et al., Recurrent nets that time and count. Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium.… [cited by applicant]
González-Pérez et al., Biomimetic triblock copolymer membrane arrays: a stable template for functional membrane proteins. Langmuir. Sep. 15, 2009;25(18):10447-50. doi: 10.1021/1a902417m. [cited by applicant]
Graves et al., Connectionist Temporal Classification: Labelling Unsegmented Sequence Data with Recurrent Neural Networks. Proceedings of the 23rd International Conference on Machine Learning, Pittsburgh, PA. 2006:369-76… [cited by applicant]
Graves et al., Speech recognition with deep recurrent neural networks. 2013 IEEE International Conference on Acoustics, Speech and Signal Processing. May 26-31, 2013. doi: 10.1109/ICASSP.2013.6638947. Epub Oct. 21, 2013… [cited by applicant]
Graves, Sequence Transduction with Recurrent Neural Networks. International Conference on Machine Learning: Representation Learning Workshop. 2012. 9 pages. [cited by applicant]
Heron et al., Simultaneous measurement of ionic current and fluorescence from single protein pores. J Am Chem Soc. Feb. 11, 2009;131(5):1652-3. doi: 10.1021/ja808128s. [cited by applicant]
Hochreiter et al., Long short-term memory. Neural Comput. Nov. 15, 1997;9(8):1735-80. [cited by applicant]
Ivanov et al., DNA tunneling detector embedded in a nanopore. Nano Lett. Jan. 12, 2011;11(1):279-85. doi: 10.1021/n1103873a. Epub Dec. 6, 2010. [cited by applicant]
Lafferty et al., Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data. Proceedings of the 18th International Conference on Machine Learning 2001. Jun. 2001. 10 pages. [cited by applicant]
Lieberman et al., Processive replication of single DNA molecules in a nanopore catalyzed by phi29 DNA polymerase. J Am Chem Soc. Dec. 22, 2010;132(50):17961-72. doi: 10.1021/ja1087612. Epub Dec. 1, 2010. [cited by applicant]
Luan et al., Base-by-base ratcheting of single stranded DNA through a solid-state nanopore. Phys Rev Lett. Jun. 11, 2010;104(23):238103. Epub Jun. 10, 2010. [cited by applicant]
McCallum et al., Maximum Entropy Markov Models for Information Extraction and Segmentation. Proceedings of the Seventeenth International Conference on Machine Learning Jun. 2000. pp. 591-598. 8 pages. [cited by applicant]
Soni et al., Synchronous optical and electrical detection of biomolecules traversing through solid-state nanopores. Rev Sci Instrum. Jan. 2010; 81(1): 014301. Epub Jan. 19, 2010. doi: 10.1063/1.3277116. [cited by applicant]
Stoddart et al., Multiple base-recognition sites in a biological nanopore: two heads are better than one. Angew Chem Int Ed Engl. 2010;49(3):556-9. doi: 10.1002/anie.200905483. [cited by applicant]
Stoddart et al., Nucleobase recognition in ssDNA at the central constriction of the alpha-hemolysin pore. Nano Lett. Sep. 8, 2010;10(9):3633-7. doi: 10.1021/n1101955a. [cited by applicant]
Stoddart et al., Single-nucleotide discrimination in immobilized DNA oligonucleotides with a biological nanopore. Proc Natl Acad Sci U S A. May 12, 2009;106(19):7702-7. doi: 10.1073/pnas.0901054106. Epub Apr. 20, 2009. [cited by applicant]
Stoiber et al., BasecRAWller: Streaming Nanopore Basecalling Directly form Raw Signal. bioRxiv Preprint. May 1, 2017. doi: http://dx.doi/org/10.11101.133058. Retrieved from https://www.biorxiv.org/content/biorxi v/early… [cited by applicant]
International Search Report and Written Opinion from Application No. PCT/GB2019/052456, mailed Dec. 13, 2019. [cited by applicant]
International Preliminary Report on Patentability for Application No. PCT/GB2019/052456, mailed Mar. 18, 2021. [cited by applicant]
Eid et al., Real-time DNA sequencing from single polymerase molecules. Science. Jan. 2, 2009; 323(5910): 133-8. doi: 10.1126/science.1162986. Epub Nov. 20, 2008. [cited by applicant]
Fuller et al., Real-time single-molecule electronic DNA sequencing by synthesis using polymer-tagged nucleotides on a nanopore array. Proc Natl Acad Sci U S A. May 10, 2016;113(19):5233-8. doi: 10.1073/pnas.1601782113. … [cited by applicant]
Haque et al., Solid-State and Biological Nanopore for Real-Time Sensing of Single Chemical and Sequencing of DNA. Nano Today. Feb. 2013;8(1):56-74. doi: 10.1016/j.nantod.2012.12.008. [cited by applicant]
Im et al., Recognition Tunneling of Canonical and Modified RNA Nucleotides for Their Identification with the Aid of Machine Learning. ACS Nano. Jul. 24, 2018; 12(7): 7067-7075. [cited by applicant]
Ni et al., DeepSignal: detecting DNA methylation state from Nanopore sequencing reads using deep-learning. Bioinformatics. Nov. 15, 2019;35(22):4586-95.doi: 10.1093/bioinformatics/btz276. [cited by applicant]
Nivala et al., Unfoldase-mediated protein translocation through an α-hemolysin nanopore. Nat Biotechnol. Mar. 2013; 31(3): 247-50. Author manuscript. 10 pages. [cited by applicant]
Rang et al., From squiggle to basepair: computational approaches for improving nanopore sequencing read accuracy. Genome Biol. Jul. 13, 2018;19(1):90. doi: 10.1186/s13059-018-1462-9. [cited by applicant]
Sedlazeck et al., Piercing the dark matter: bioinformatics of long-range sequencing and mapping. Nat Rev Genet. Jun. 2018; 19(6): 329-346. [cited by applicant]
Simpson et al., Detecting DNA cytosine methylation using nanopore sequencing. Nat Methods. Apr. 2017;14(4):407-410. doi: 10.1038/nmeth.4184. Epub Feb. 20, 2017. [cited by applicant]
Soltau et al., Neural speech recognizer: Acoustic-to-word LSTM model for large vocabulary speech recognition. arXiv preprint arXiv:1610.09975. Oct. 31, 2016. [cited by applicant]
Stoiber et al., De novo Identification of DNA Modifications Enabled by Genome-Guided Nanopore Signal Processing. BioRxiv Preprint. Apr. 10, 2017; doi: https://doi.org/10.1101/094672. Accessible at https://www.biorxiv.or… [cited by applicant]
Teng et al., Chiron: translating nanopore raw signal directly into nucleotide sequence using deep learning. Gigascience. May 1, 2018; 7(5): 1-9. [cited by applicant]
Third Party Observations for Application No. EP 18723942.1, mailed May 3, 2023. [cited by applicant]
Wang et al., Solid-State Nanopore Analysis of Diverse DNA Base Modifications Using a Modular Enzymatic Labeling Process. Nano Lett. Nov. 8, 2017; 17(11): 7110-7116. doi: 10.1021/acs.nanolett.7b03911. Epub Oct. 5, 2017. [cited by applicant]
Extended European Search Report for Application No. EP 24170727.2, mailed Jul. 25, 2024. [cited by applicant]
[No Author Listed] Oxford Nanopore Technologies: “Basecalling overview”, Aug. 22, 2016 (Aug. 22, 2016), XP093186903, Retrieved from the Internet: URL:https://community.nanoporetech.com/technical_documents/data-analysis/… [cited by applicant]
Ansorge, Next-generation DNA sequencing techniques. N Biotechnol. Apr. 2009;25(4):195-203. doi: 10.1016/j.nbt.2008.12.009. Epub Feb. 3, 2009. [cited by applicant]
Trigeorgis et al., Adieu features? end-to-end speech emotion recognition using a deep convolutional recurrent network. 2016 IEEE international conference on acoustics, speech and signal processing (ICASSP) Mar. 20, 2016… [cited by applicant]
EP 24170727.2, Jul. 25, 2024, Extended European Search Report. [cited by applicant]