How nanopore sequencing works: pore architectures and signal physics

Nanopore DNA sequencing identifies nucleotide sequences by measuring ionic current fluctuations as single DNA or RNA molecules translocate through nano-scale protein or solid-state pores. As each nucleotide — or k-mer window — passes through the aperture, it partially occludes the ionic current, producing a characteristic electrical signal from which base identity and, increasingly, base modification status are inferred.

Among retrieved results, three principal pore architectures define the competitive landscape. Biological (protein) nanopores — engineered protein channels such as alpha-hemolysin (α-HL) and MspA mutants inserted into lipid bilayers — are the dominant commercial approach. The University of Southampton’s molecular dynamics analysis of α-HL’s nucleotide capture region illustrates ongoing refinement of the geometry governing translocation speed and signal differentiation, while Southeast University’s comparative analysis of MspA mutants demonstrates the engineering of pore geometry for improved discrimination.

Solid-state nanopores — lithographically or chemically fabricated pores in silicon nitride, graphene, or related materials — offer tunability and CMOS integration potential. According to KTH Royal Institute of Technology‘s computational study, persistent challenges for solid-state variants include translocation speed control and signal noise. A third emerging class, DNA-nanostructure hybrid pores, combines the addressability of DNA origami with solid-state substrates, as documented by Aalto University’s review of DNA nanopore techniques used as customised gates in lipid membranes and nanocapillaries.

The dominant commercial platform across the dataset is Oxford Nanopore Technologies (ONT), whose MinION, GridION, and PromethION devices are referenced across the vast majority of retrieved results. According to WIPO‘s global sequencing technology filings, the nanopore category has seen accelerating patent activity since 2017, with hardware miniaturisation and signal processing algorithms representing the two fastest-growing sub-categories.

From 38% error rates to 96.5% accuracy: the four-phase innovation timeline

The nanopore sequencing field divides into four recognisable innovation phases based on publication dates across the 80+ retrieved records, each defined by a distinct technical challenge and resolution.

2010–2014 — Foundational Infrastructure: Early records focus on design theory, translocation physics, and preliminary platform validation. Science for Life Laboratory’s 2010 work on binary DNA encoding for nanopores and KTH’s 2012 overview of sequencing generations bracket this era. The University of Birmingham’s 2014 first reference bacterial genome dataset generated on the MinION marks first-generation hardware commercialisation under the MinION Access Program (MAP).

2015–2016 — Performance Benchmarking and Early Error Characterisation: A cluster of results from University of Exeter, UC Santa Cruz, Cold Spring Harbor Laboratory, and Brown University rigorously benchmarks error rates, read length distributions, and assembly quality. The University of Exeter’s assessment estimated a post-basecalling error rate of 38.2% on early chemistry. UC Santa Cruz demonstrated greater than 99% mapping rates for high-quality 2D reads. The Nanopore-CMOS interface study by York University (2016) pointed toward hardware integration for signal amplification at scale.

2017–2020 — Deep Learning Basecalling and Targeted Sequencing: The University of Queensland’s Chiron paper (2017) introduced the first end-to-end deep learning basecaller achieving more than 2,000 bases per second without segmentation. This era also witnessed the first human genome assembly from nanopore ultra-long reads (UC Santa Cruz, 2017), GPU-accelerated adaptive sampling at gigabase scale (University of Nottingham, 2020), and mobile smartphone-based analysis pipelines (Garvan Institute, 2020).

2021–2024 — Clinical Deployment, RNA Sequencing, and Competitive Entry: The most recent filings and publications reflect mature clinical applications, direct RNA sequencing, adaptive sampling for pathogen enrichment, and a notable patent from Illumina (2024, EP) signalling competitive platform entry.

“Accuracy improved from ~65% on early ONT reads to 96.5% over a five-year window — a trajectory driven entirely by pore chemistry refinement and the shift from HMM-based to deep learning basecalling.”

Four technology clusters driving the nanopore patent landscape

The retrieved dataset organises into four distinct technology clusters, each representing a separable IP and R&D arena with different maturity profiles and competitive dynamics.

Cluster 1: Biological pore strand sequencing

The dominant approach across the dataset. A motor protein (e.g., helicase) ratchets single-stranded DNA through a protein nanopore — MspA, CsgG, or α-HL — embedded in a lipid bilayer, while a patch-clamp amplifier measures picoampere-level current changes. Signal segments corresponding to 5- or 6-mer windows are classified by a trained model. According to EMBL-EBI‘s genomics data resources, biological pore sequencing now accounts for the majority of long-read sequencing datasets deposited in public archives.

Cluster 2: Deep learning basecalling and signal processing

Raw electrical signals (“squiggles”) are converted to base sequences using recurrent neural networks (RNNs), convolutional networks, and residual networks. Chiron (2017, University of Queensland) pioneered end-to-end translation without segmentation, achieving more than 2,000 bases per second. DeepNano-blitz (2020, Comenius University) optimised for CPU-only real-time deployment. RawHash (2023, ETH Zurich) introduced hash-based raw signal mapping for large genomes under Read Until — bypassing basecalling entirely to support real-time enrichment decisions at human-genome scale.

Cluster 3: Adaptive and targeted sequencing (Read Until)

A uniquely nanopore-native capability: the sequencer can reverse voltage to eject unwanted molecules from a pore mid-translocation, enabling real-time computational enrichment or depletion without prior PCR. University of Nottingham’s 2020 study demonstrated GPU-accelerated basecalling enabling mapping against gigabase references in real time. Johns Hopkins University’s UNCALLED (2020) demonstrated targeted nanopore sequencing by real-time mapping of raw electrical signal. Southern University of Science and Technology of China’s ultra-sensitive antimicrobial resistance workflow (2022) further demonstrated adaptive sampling for host DNA depletion in complex clinical matrices.

Cluster 4: Solid-state, hybrid, and alternative pore architectures

Solid-state nanopores (silicon nitride, graphene) and DNA-origami hybrid pores offer tunability, durability, and integration with CMOS electronics. Osmylated DNA sequencing (binary chemical encoding, Yenos Analytical LLC, 2015) and Nanopore-Induced Phase-Shift Sequencing (NIPSS) represent alternative signal encoding strategies. Illumina’s 2024 EP patent on modified electrolytes in nanopore cis/trans wells signals corporate entry into pore chemistry innovation — a direct challenge to ONT’s biological pore model.

Application domains: from clinical diagnostics to the International Space Station

Nanopore DNA sequencing has been deployed across six distinct application domains, with clinical infectious disease diagnostics representing the largest and most commercially mature cluster in the retrieved dataset.

Clinical infectious disease diagnostics

Nanopore Targeted Sequencing (NTS) demonstrated advantages over culture in detection speed and range, according to Wuhan University’s 2022 clinical performance study. Direct RNA sequencing of SARS-CoV-2 from oropharyngeal swabs without cDNA conversion was demonstrated by CNR-ICAR (Italy, 2022). Anthrax whole-genome characterisation was completed within hours of isolate receipt by the CDC (2020). Ebola field diagnostics were demonstrated from a Liberia outbreak laboratory by the CDC (2016).

Oncology and precision medicine

Copy number alteration inference in cancer using short-molecule nanopore sequencing was validated at Memorial Sloan Kettering (2020). Shallow nanopore RNA sequencing for transcriptome profiling of tumours — with a five-day clinical turnaround — was demonstrated across four cancer types (2022). Rolling-circle amplification strategies enable accurate sequencing of ultra-short cell-free DNA fragments relevant to liquid biopsy, developed at Stanford University (2019).

Epigenomics and base modification detection

Nanopore sequencing’s native ability to detect 5-methylcytosine, 5-hydroxymethylcytosine, and other base modifications directly from native DNA — without bisulfite conversion — positions it uniquely for epigenetic epidemiology. The University of Exeter benchmarked nanopore methylation calls against Illumina EPIC arrays using CRISPR-Cas9 targeted enrichment (2022). Cambridge’s DNAscent v2 applies residual neural networks to detect BrdU base analogues at single-nucleotide resolution for replication fork mapping (2021).

Environmental monitoring and metagenomics

The ONT MinION enables sequencing in resource-limited environments. Water research applications — including field sequencing aboard diving vessels and oceanographic ships — are documented by Ardhi University (Tanzania, 2022). Full-length 16S rRNA sequencing for microbiome characterisation achieves species-level classification at 0.1% abundance, according to the Genome Institute of Singapore (2016).

Space and extreme environments

The MinION was operated successfully on the International Space Station and in parabolic microgravity flights, generating more than 80,000 reads with 85–90% mean 2D accuracy, according to a UCSF study (2016). This positions nanopore sequencing as the only viable real-time sequencing technology for space biology and astrobiology.

DNA data storage

Stanford University’s 2021 theoretical work on coding schemes for nanopore-read DNA storage addresses capacity bounds for an abstracted nanopore channel model. North China Electric Power University reviewed nanopore detection as the read-out mechanism for DNA computing, storage, and self-assembly applications (2022).

Five emerging directions shaping nanopore R&D through 2026

The most recent filings and publications (2022–2024) across the dataset reveal five convergent directions that define where nanopore sequencing innovation is heading over the next two to three years.

1. Adaptive sampling at clinical scale. NASCarD (2023) demonstrates PCR-free whole-genome sequencing of SARS-CoV-2 with more than 100× enrichment over shotgun metatranscriptomics using adaptive sampling combined with carrier DNA, directly in clinical diagnostic pipelines.

2. Real-time raw signal analysis. ETH Zurich’s RawHash (2023) enables hash-based similarity search directly on raw electrical signals — bypassing basecalling entirely — to support Read Until at scales covering the full human genome. This represents a fundamental shift in how the computational layer interacts with sequencing hardware.

3. Direct RNA sequencing and transcriptomics. Multiple 2021–2022 results describe maturation of direct RNA sequencing (without cDNA conversion), including NanopoReaTA for real-time differential transcriptome analysis (2022) and MasterOfPores for ONT direct RNA datasets from the Centre for Genomic Regulation, Barcelona (2020). This enables simultaneous sequencing of sequence, RNA modifications, and polyA tail length from a single molecule.

4. Competitive commercial platform entry. Illumina’s active 2024 EP patent covering modified electrolyte and gel-state polyelectrolyte configurations in nanopore cis/trans wells is the most significant IP signal in the dataset. This suggests Illumina is building a proprietary nanopore pore-chemistry IP position as it develops competitive platforms, according to EPO patent records.

5. Portable computation and mobile genomics. Smartphone-native analysis (Genopo for Android, F5N) and CPU-only basecalling (DeepNano-blitz) converge with sub-30-minute sample-to-answer workflows. The Garvan Institute demonstrated SARS-CoV-2 genome assembly on a smartphone (2020). This trajectory points toward fully portable, offline-capable sequencing and analysis units suitable for point-of-care and low-resource settings.

“Illumina’s 2024 EP active patent on modified electrolytes in nanopore cis/trans wells is the most significant IP signal in the dataset — marking the first credible challenge to ONT’s biological pore chemistry dominance.”

IP strategy: white space, contested arenas, and first-mover opportunities

The strategic implications of this landscape assessment translate directly into five IP and R&D positioning priorities for teams operating in or adjacent to the nanopore sequencing space.

Pore chemistry and electrolyte engineering represent the most significant IP white space. Illumina’s 2024 EP active patent on modified electrolytes in nanopore wells signals that this layer — long dominated by ONT’s proprietary biological pore configurations — is becoming a contested IP arena. R&D teams should evaluate freedom-to-operate and filing opportunities in pore-membrane interface chemistry, particularly for solid-state and hybrid pore variants.

Deep learning basecalling is the field’s fastest-moving software frontier. The shift from HMM-based to CNN/RNN-based basecalling (Chiron 2017 → DeepNano-blitz 2020 → GPU-accelerated models 2022+) and now to raw-signal hashing (RawHash 2023) creates continuous IP layering in signal processing algorithms. Teams building AI-native sequencing infrastructure should monitor patent filings around real-time k-mer encoding and transformer-based basecalling architectures not yet appearing prominently in this dataset.

Clinical infectious disease is the near-term highest-value application domain. The concentration of results in pathogen detection, AMR characterisation, and outbreak response — with demonstrated sub-four-hour sample-to-result times — indicates a maturing clinical market. IP strategists should map existing freedom-to-operate around multiplex amplicon sequencing workflows, metagenomics pipelines, and adaptive sampling for host DNA depletion in clinical matrices.

Epigenomics represents an underexploited differentiation vector. Nanopore’s ability to detect 5-methylcytosine, 5-hydroxymethylcytosine, and other base modifications from native DNA without chemical conversion is unique among sequencing platforms. The benchmarking work by University of Exeter (2022) and Cambridge’s DNAscent v2 (2021) signal early-stage validation that has not yet translated into a dense patent thicket — a potential first-mover opportunity.

The portable computation stack is approaching convergence with the portable hardware stack. As smartphone-based analysis, CPU-only real-time basecalling, and sub-30-minute complete workflows converge with the MinION/Flongle hardware form factor, the integrated point-of-care genomics device becomes technically feasible. Product developers should evaluate the regulatory and reimbursement pathway implications of this convergence, particularly in infectious disease diagnostics in low-resource settings. For context on regulatory frameworks, the FDA‘s guidance on next-generation sequencing-based diagnostics provides the relevant clinical validation framework in the US market.

Geographic distribution of innovation is broad. The United States leads by result count — with UC Santa Cruz, Stanford, Johns Hopkins, Columbia, Harvard, Memorial Sloan Kettering, CDC, UCSF, and Intel all contributing — but China (Chinese Academy of Sciences, Southeast University, Wuhan University, Southern University of Science and Technology) and the United Kingdom (Birmingham, Exeter, Nottingham, Cambridge) represent substantial and growing innovation hubs. European contributions from ETH Zurich, CEA/Genoscope, and the Helmholtz Institute round out the landscape. Teams building global IP portfolios should ensure filing strategies account for all three jurisdictions.