Nanopore Full-Length cDNA Sequencing — High-Throughput Full-Length Transcript Isoform Discovery & Quantification

Full-length splice variant analysis, fusion gene detection, and isoform-level expression profiling by Oxford Nanopore cDNA sequencing on PromethION — PCR-cDNA and direct cDNA workflows supported

CD Genomics provides Nanopore Full-Length cDNA Sequencing on the PromethION platform, delivering 50–100+ million reads per flow cell that capture complete transcript isoforms as single contiguous reads. Using both PCR-cDNA and direct cDNA workflows, our service enables full-length isoform discovery, splice variant detection, fusion gene identification, and isoform-level quantification across any eukaryotic species — without the computational reconstruction required by short-read RNA-seq methods.

The transcriptome is not defined by gene expression levels alone — it is defined by which isoforms are expressed, in what proportions, and how splice variation, alternative promoter usage, and alternative polyadenylation reshape the functional landscape of the cell. Short-read RNA sequencing, which fragments transcripts into 50–300 bp pieces before sequencing, observes the transcriptome through fragmented windows and relies on computational inference to reconstruct full-length isoform architectures from partial evidence. As transcript complexity increases — with multi-exon skipping, mutually exclusive exons, intron retention events, and complex alternative splicing patterns — the inference problem becomes increasingly ill-posed, and short-read data systematically underestimates isoform diversity.

CD Genomics offers Nanopore Full-Length cDNA Sequencing — a high-throughput long-read RNA sequencing service that captures complete transcript isoforms as single contiguous reads. Using Oxford Nanopore Technologies (ONT) PromethION and GridION platforms with both PCR-cDNA and direct cDNA library preparation workflows, our service delivers full-length transcript sequences spanning from the 5′ cap region to the 3′ poly(A) tail, enabling direct observation of splice isoform architecture, isoform-level quantification across experimental conditions, fusion transcript discovery, and alternative splicing analysis at single-molecule resolution without computational reconstruction.

Why Nanopore Full-Length cDNA Sequencing — Service Highlights

• High-throughput full-length transcript sequencing — Nanopore cDNA sequencing delivers 50–100+ million reads per PromethION flow cell, 5–10 times more than direct RNA sequencing and competitive with short-read throughput. This depth enables detection of low-abundance isoforms, rare splicing events, and transcripts expressed at a few copies per cell that would be missed by lower-throughput long-read methods
• Two complementary library preparation options — PCR-cDNA (ONT SQK-PCS111) provides maximum throughput with template-switching-based 5′ cap capture for comprehensive isoform discovery, while direct cDNA (ONT SQK-DCS111) eliminates PCR amplification bias and preserves strand-of-origin information for strand-aware transcript analysis. We recommend the optimal workflow based on sample availability, project throughput requirements, and the importance of strand information
• Full-length transcript connectivity without computational reconstruction — Each cDNA read represents a complete transcript molecule spanning from 5′ cap to 3′ poly(A) tail. Splice isoform architecture, alternative promoter usage, alternative polyadenylation (APA) site selection, and complex alternative splicing patterns are observed directly rather than inferred from fragmented reads — providing definitive evidence for transcript structures that short-read methods can only approximate
• Validated single-cell and low-input compatibility — We offer established protocols for sequencing 10x Genomics single-cell cDNA libraries on the ONT platform, enabling single-cell full-length transcript isoform discovery with demonstrated performance across multiple 10x chemistries. Our service also supports bulk RNA inputs as low as 100 ng total RNA, extending full-length transcript analysis to sample-limited experimental contexts including biopsies, laser-capture microdissected tissue, and rare cell populations
• Accurate isoform quantification and differential analysis — Dedicated bioinformatics pipelines (bambu, Isosceles, Flair) provide isoform-level expression quantification, differential transcript usage (DTU) analysis, and splice junction quantification from full-length cDNA reads. Isoform expression estimates are robust across expression ranges, enabling confident identification of condition-specific isoform switching events that drive functional changes in disease states and developmental transitions
• Integrated multi-platform RNA analysis ecosystem — Our Nanopore cDNA sequencing service is part of a comprehensive RNA analysis portfolio. Combined with Nanopore Direct RNA Sequencing for epitranscriptome analysis on matching samples, PacBio RNA Sequencing for HiFi-accuracy isoform validation, and dedicated bioinformatics support through Oxford Nanopore Sequencing Data Analysis, researchers can design multi-platform strategies that leverage the strengths of each technology

Full-Length cDNA Sequencing Reveals Transcript Isoform Architecture That Short-Read Methods Cannot Resolve

The human transcriptome is estimated to produce over 200,000 distinct transcript isoforms from approximately 20,000 protein-coding genes, with the majority of multi-exon genes expressing multiple splice variants that differ in their exon composition, untranslated region architecture, and coding potential. These isoforms are not functionally equivalent. Alternative splicing can alter protein domain composition, change subcellular localization signals, introduce or remove regulatory elements in untranslated regions, and produce transcripts targeted for nonsense-mediated decay. The choice between isoforms is cell-type-specific, developmentally regulated, and frequently dysregulated in disease — yet short-read RNA-seq, the most widely used transcriptome profiling method, provides only indirect evidence for these isoform-level events.

Full-length cDNA sequencing addresses this gap by observing complete transcript isoforms directly. Unlike short-read methods that fragment RNA before sequencing and then computationally reassemble the fragments into inferred transcript models, full-length cDNA sequencing captures each transcript molecule in its entirety as a single sequencing read. The isoform structure — which exons are included, which splice sites are used, where the transcript starts and ends — is read directly from the sequencing data, not inferred from fragment coverage patterns. This distinction is critical for accurate isoform-level analysis: studies consistently show that short-read methods systematically underestimate isoform diversity, miss rare isoforms, and produce ambiguous isoform assignments for genes with complex splicing patterns.

Nanopore cDNA sequencing occupies a specific position in the long-read RNA sequencing landscape. It provides the highest throughput among full-length transcript methods (50–100+ million reads per flow cell), making it the method of choice when the biological question requires deep coverage of the transcriptome for comprehensive isoform discovery and reliable quantification of differentially expressed isoforms. Unlike our Nanopore Direct RNA Sequencing service, cDNA sequencing requires reverse transcription and (in the PCR-cDNA workflow) PCR amplification, which means RNA base modification information is not preserved. The trade-off is throughput: where direct RNA sequencing provides 5–20 million reads per flow cell with full modification information, cDNA sequencing delivers an order of magnitude more reads for projects where transcript structure and expression, not modifications, are the primary question.

Nanopore Full-Length cDNA Sequencing Delivers High-Throughput Isoform Discovery with Single-Molecule Resolution

Nanopore Full-Length cDNA Sequencing Supports Isoform Discovery Across Cancer, Single-Cell, Neurobiology, and Plant Research

The PCR-cDNA workflow's deep sequencing depth (50–100+ million reads per flow cell) enables detection of low-abundance isoforms, while the direct cDNA workflow's strand-aware protocol provides accurate annotation of antisense transcripts and non-coding RNA isoforms — a capability not available from most other long-read RNA sequencing services.

Comprehensive Transcript Isoform Discovery and Annotation

• Genome-wide identification of full-length transcript isoforms across any eukaryotic species, including discovery of novel isoforms not present in reference annotations, alternative promoter usage events, alternative splicing patterns, and alternative polyadenylation (APA) site selection — providing definitive full-length read evidence for each transcript structure
• De novo transcriptome annotation for non-model organisms where reference genome annotations are incomplete or absent, generating comprehensive isoform catalogs with experimentally determined 5′ and 3′ boundaries, exon-intron structures, and splice junction evidence from full-length reads
• Integration with short-read RNA-seq data to resolve isoform ambiguity: short-read data provides deep coverage for expression quantification, while full-length cDNA reads resolve the isoform structures that short reads cannot distinguish, creating hybrid assemblies with both breadth and structural resolution

Cancer Transcriptomics and Fusion Gene Discovery

• Detection of cancer-specific transcript isoforms including oncogenic splice variants, aberrant exon inclusion events, and tumor-specific exon skipping patterns from tumor RNA samples, with full-length read evidence for the complete isoform structure rather than inference from partial junction reads
• Fusion transcript discovery and characterization using full-length reads that span fusion breakpoints, providing exact fusion junction sequences, isoform structures of fusion transcripts, and expression levels of fusion isoforms — including complex fusion events involving >2 genes or intragenic rearrangements that are difficult to resolve with short-read methods
• Isoform-level analysis of tumor suppressors and oncogenes, where different splice isoforms can have opposing functions (e.g., pro-apoptotic vs anti-apoptotic BCL-X isoforms, full-length vs truncated p53 isoforms), enabling isoform-resolved functional interpretation of cancer transcriptome data
• Integration with Cancer Research applications for comprehensive transcriptomic characterization of tumor samples, including paired analysis with direct RNA sequencing for simultaneous isoform and modification analysis in cancer pathways

Single-Cell Full-Length Transcriptomics

• Single-cell isoform discovery from 10x Genomics single-cell cDNA libraries sequenced on ONT platforms, enabling cell-type-resolved analysis of alternative splicing, isoform expression heterogeneity, and cell-state-specific transcript usage at the single-cell level
• Integration of single-cell full-length transcript data with single-cell gene expression data to link cell-type-specific isoform usage with cell identity, differentiation trajectories, and functional states — revealing splicing regulation at cellular resolution that is masked in bulk transcriptome analysis
• Detection of single-cell-level isoform switching during development, differentiation, and disease progression, with dedicated bioinformatics pipelines (Isosceles, Bambu-Clump) optimized for the specific challenges of single-cell long-read data including lower per-cell coverage and higher technical noise

Neurobiology and Neurodegenerative Disease Transcriptomics

• Profiling of neuron-specific alternative splicing events, including microexon inclusion, alternative tandem exon usage, and RNA editing-dependent isoform diversity in nervous system tissues, where alternative splicing regulation is particularly complex and functionally significant
• Isoform-level analysis of genes implicated in neurodegenerative diseases (APP, MAPT/Tau, SNCA, TARDBP/TDP-43, C9orf72) where specific splice isoforms are directly linked to disease pathogenesis, enabling isoform-resolved characterization of disease-associated transcript changes
• Characterization of antisense transcripts and non-coding RNA isoforms in neuronal tissues, leveraging the strand-aware analysis capability of direct cDNA sequencing for accurate annotation of natural antisense transcripts and long non-coding RNA isoforms expressed in the brain

Plant and Non-Model Organism Transcriptomics

• Full-length transcriptome characterization for plant species, agricultural crops, and non-model organisms where reference genome annotations are incomplete or unavailable, providing experimentally determined transcript structures that enable accurate gene model annotation, splice variant cataloging, and functional genomic analysis
• Isoform-level analysis of stress-responsive transcript regulation in plants, including identification of condition-specific splice isoforms induced by drought, salinity, temperature stress, and pathogen challenge — supporting crop improvement and stress tolerance research
• Comparative transcriptomics across species, populations, or cultivars to identify conserved and species-specific isoform usage patterns, alternative splicing events associated with domestication or adaptation, and evolutionary changes in transcript structure across the tree of life

From RNA to Full-Length Reads: The Nanopore cDNA Sequencing Workflow

1. RNA Extraction and Quality Assessment

Total RNA is extracted from the sample using methods optimized for RNA integrity and purity. RNA quantity and quality are assessed by microfluidic electrophoresis (RIN ≥ 8 recommended for standard protocols). Poly(A) selection using oligo-dT beads is performed for mRNA-focused studies; rRNA depletion is available for total RNA analysis including non-coding transcripts. DNase treatment eliminates genomic DNA contamination. For single-cell cDNA sequencing projects, cDNA is generated from 10x Genomics single-cell libraries using the 10x pipeline and then converted to ONT-compatible libraries.

2. Full-Length cDNA Synthesis

For the PCR-cDNA workflow (ONT SQK-PCS111), reverse transcription is performed using an oligo(dT) primer that captures the 3′ poly(A) tail. The template-switching oligonucleotide (TSO) captures the 5′ cap structure of full-length mRNAs, enabling the reverse transcriptase to complete full-length first-strand cDNA synthesis. The resulting cDNA contains the complete transcript sequence from the 5′ cap region through the coding sequence to the 3′ poly(A) tail, with adapters at both ends for subsequent PCR amplification and sequencing. For the direct cDNA workflow (ONT SQK-DCS111), the oligo(dT) primer includes the motor protein adapter sequence, enabling direct sequencing of the first-strand cDNA without PCR amplification. Reverse transcription is performed with a thermostable RT enzyme optimized for full-length cDNA synthesis, minimizing premature termination at structured RNA regions.

Figure 1. Nanopore Full-Length cDNA Sequencing workflow: from RNA extraction and quality assessment through full-length cDNA synthesis (PCR-cDNA or direct cDNA library preparation) and Oxford Nanopore sequencing on PromethION or GridION platforms.

3. Library Preparation and Adapter Ligation

For PCR-cDNA: The full-length cDNA is PCR-amplified using primers specific to the adapters introduced during reverse transcription, with the number of PCR cycles optimized to balance yield against amplification bias (typically 12–18 cycles). Sequencing adapters containing the motor protein and tether are ligated to the amplified cDNA. For direct cDNA: The first-strand cDNA, already containing the motor protein adapter incorporated during reverse transcription, is directly prepared for sequencing without amplification. In both workflows, the library is purified to remove short fragments, unligated adapters, and other contaminants before loading onto the flow cell.

4. Nanopore Sequencing on PromethION or GridION

The prepared library is loaded onto a PromethION or GridION flow cell. The motor protein bound to the cDNA adapter processively unwinds and translocates the cDNA molecule through the nanopore at approximately 400–450 bases per second. As each nucleotide passes through the pore, it modulates the ionic current in a sequence-dependent manner. The current is measured at kHz frequency, generating a continuous signal trace that encodes the complete nucleotide sequence of the cDNA molecule. Basecalling is performed in real time using ONT Dorado with super-accuracy (SUP) models, producing high-quality FASTQ reads for downstream analysis. The sequencing run can be monitored in real time, with run termination decisions based on accumulated data yield, target coverage, and project-specific quality metrics.

5. Bioinformatics Analysis for Isoform Discovery and Quantification

Basecalled reads are filtered by quality score and read length, then aligned to the reference genome or transcriptome using splice-aware aligners (minimap2 with splice parameters, or uLTRA for more sensitive alignment of full-length transcripts). Downstream analysis includes: isoform detection and quantification using dedicated long-read tools (bambu for reference-based isoform discovery and quantification, Isosceles for single-cell-resolution analysis, Flair for full-length isoform assembly and quantification, or isoform quantification tool for comprehensive isoform analysis with splice junction support), differential transcript usage (DTU) analysis (DRIMSeq or satuRn), splice junction quantification and alternative splicing event analysis (SUPPA2 or MAJIQ), fusion transcript detection (JAFFAL or LongGF), and functional annotation of novel isoforms including coding potential assessment, conserved domain identification, and comparison with reference annotations.

Sample Input Requirements for PCR-cDNA and Direct cDNA Workflows

Case Study: Accurate Long-Read Transcript Discovery and Quantification at Single-Cell, Pseudo-Bulk and Bulk Resolution with Isosceles

Kabza M, Ritter A, Byrne A, Sereti K, Le D, Stephenson W, Sterne-Weiler T. Accurate long-read transcript discovery and quantification at single-cell, pseudo-bulk and bulk resolution with Isosceles. Nature Communications. 2024;15:7316. (CC BY 4.0)

1. Background

Long-read RNA sequencing technologies, including Nanopore cDNA sequencing, offer the ability to sequence full-length transcript isoforms, but the computational analysis of long-read data presents unique challenges. Technical noise, variable read quality, and the complexity of isoform-level quantification — particularly at the single-cell level — require specialized computational methods that can distinguish true biological isoform diversity from technical artifacts. Existing tools (Bambu, isoform quantification tool, ESPRESSO) each have limitations in sensitivity and accuracy, especially for single-cell data where per-cell coverage is limited. The authors developed Isosceles, a computational toolkit designed to address these challenges by using acyclic splice-graph representations of gene structure combined with an Expectation-Maximization (EM) algorithm for accurate isoform detection and quantification from Nanopore long-read RNA sequencing data.

2. Methods

Isosceles was developed using a splice-graph-based approach in which gene structure is represented as a directed acyclic graph, with exons as vertices and splice junctions as edges. Both known transcripts from reference annotations and de novo transcripts identified from long-read data are incorporated into the graph. Transcript compatibility counts (TCCs) are calculated by aligning each long read to the graph and determining which transcripts are compatible with the observed splice junctions. An EM algorithm is then applied to the TCCs to estimate transcript-level expression. The method was benchmarked using synthetic RNA spike-in data with known isoform composition, as well as bulk and single-cell Nanopore cDNA sequencing data from human cell lines and mouse brain. Performance was compared against Bambu, isoform quantification tool, and ESPRESSO across multiple metrics including isoform detection sensitivity, quantification accuracy, and computational efficiency.

3. Results

Figure 2. Isosceles architecture and key results: splice-graph-based transcript discovery and quantification from Nanopore long-read cDNA sequencing data, with validation across single-cell, pseudo-bulk, and bulk resolutions. From Kabza M, Ritter A, Byrne A, et al. (2024, Nature Communications, CC BY 4.0).

Key Findings

• Superior isoform detection and quantification accuracy across resolutions — Isosceles achieved a Spearman correlation of 0.96 for annotated transcript quantification versus ground-truth data, outperforming Bambu (0.92), isoform quantification tool (0.88), and ESPRESSO (0.88). The median quantification error was reduced by 21% compared to the next best method (Bambu), and at single-cell resolution, Isosceles maintained 24–43% lower mean relative difference than all other tested methods
• De novo transcript discovery with controlled false discovery rates — The method's de novo detection mode identified novel transcripts with low false discovery rates (~4.3%), enabling confident discovery of previously unannotated isoforms from Nanopore cDNA sequencing data without sacrificing specificity
• Biological discovery in single-cell mouse brain data — Reanalysis of 951 single-cell Nanopore transcriptomes from mouse E18 brain using Isosceles resolved two distinct glutamatergic neuron differentiation trajectories from the developing cortex, where the original analysis had identified only one trajectory. The analysis identified 25 alternative splicing events changing within these trajectories and 21 splice events differing between the two trajectories, demonstrating the biological resolution gained from improved isoform quantification at single-cell level
• Robust performance across data types and coverages — Isosceles maintained consistent performance across bulk RNA, pseudo-bulk, and single-cell resolutions, with accuracy improvements most pronounced at lower coverage levels typical of single-cell data — demonstrating that the method is particularly well-suited for projects combining single-cell and bulk long-read transcriptome analysis
• Open-source availability enabling community adoption — The Isosceles toolkit is freely available as an open-source R/Bioconductor package, providing the research community with a validated computational resource for Nanopore full-length cDNA sequencing data analysis and supporting reproducible transcript discovery and quantification workflows

4. Conclusions

This study provides a validated computational framework for accurate transcript discovery and quantification from Nanopore full-length cDNA sequencing data, with demonstrated performance across bulk, pseudo-bulk, and single-cell resolutions. The Isosceles toolkit addresses a critical need in the long-read transcriptomics field — the challenge of accurate isoform-level quantification from noisy long-read data — and achieves significant improvements over existing methods in both sensitivity and accuracy. The biological discoveries enabled by reanalysis of mouse brain single-cell data demonstrate that improved computational methods can extract additional biological insight from existing Nanopore cDNA sequencing datasets. These findings validate our bioinformatics pipeline design for full-length cDNA sequencing projects, which includes Isosceles as a key component of the Advanced analysis package for single-cell isoform analysis and demonstrates the value of ongoing computational method development for extracting maximal biological information from Nanopore full-length cDNA sequencing data.

When to Choose Nanopore Full-Length cDNA Sequencing — and When Alternative Methods May Be More Suitable

Choose Nanopore full-length cDNA sequencing when:

• Comprehensive isoform discovery is the primary research goal — The high throughput of PCR-cDNA workflow (50–100+ million reads per flow cell) enables detection of low-abundance isoforms, rare splicing events, and transcripts expressed at a few copies per cell that lower-throughput long-read methods would miss
• Your project requires both discovery and quantification in a single workflow — Full-length cDNA reads provide both isoform structure and expression data, eliminating the need for separate short-read quantification experiments and enabling direct correlation of isoform structure with expression level
• You are working with non-model organisms or incomplete annotations — Full-length transcript sequencing generates experimentally determined isoform structures with verified 5′ and 3′ boundaries, providing definitive transcript evidence for de novo genome annotation without reliance on reference annotations
• Fusion transcript or novel transcript discovery is a stated objective — Full-length reads that span fusion breakpoints provide exact junction sequences and isoform structures of fusion transcripts, including complex events involving more than two genes that are difficult to resolve with short-read methods
• Single-cell isoform analysis is required — Our validated protocols for 10x Genomics single-cell cDNA libraries on ONT platforms enable cell-type-resolved isoform discovery, with dedicated bioinformatics pipelines optimized for single-cell long-read data

Consider alternative methods when:

• RNA modification detection is the primary objective — Nanopore cDNA sequencing does not preserve base modifications. Use Nanopore Direct RNA Sequencing when m⁶A, m⁵C, pseudouridine, or poly(A) tail length information is required
• Maximum single-base accuracy is needed for specific isoforms — For validation of specific isoform sequences or de novo transcript annotation requiring consensus accuracy, PacBio RNA Sequencing (Iso-Seq with Kinnex) provides Q30+ circular consensus accuracy that complements Nanopore cDNA throughput
• Only gene-level expression data is needed — For projects where gene-level differential expression (not isoform-level) is the only question, short-read RNA-seq remains more cost-effective per sample at scale
• RNA input is severely limited (below 50 ng) — While we support low-input protocols, samples with less than 50 ng total RNA or highly degraded RNA (RIN <5) may not yield sufficient full-length cDNA for comprehensive analysis

CD Genomics provides free project consultation to help determine the optimal RNA sequencing strategy for your specific research question. Contact our scientists to discuss your project requirements.

Full-Length cDNA Sequencing Data Deliverables Include Isoform Annotations, Expression Matrices, and Integrated Reports

In addition to standard FASTQ, BAM, and expression matrix deliverables, our Advanced package provides optional multi-platform integration reports that correlate cDNA isoform data with direct RNA modification analysis or PacBio HiFi isoform validation from the same biological samples.

1. Basecalled FASTQ files from Dorado super-accuracy (SUP) basecalling with RNA-specific models, including per-read quality metrics, adapter-trimmed sequences, and read length distributions — suitable for downstream analysis and public data deposition in SRA or GEO

2. Genome-aligned BAM files with splice-aware alignments (minimap2), transcript isoform annotations, and splice junction labels — ready for visualization in IGV or UCSC Genome Browser showing full-length read coverage across gene models with isoform-specific read assignments

3. Transcript isoform annotation and quantification report including full-length isoform structures with exon-intron boundaries, isoform-level expression matrices (TPM/read counts) for all detected isoforms, isoform novelty classification (known/novice in catalog/novel not in catalog), and comparison with reference annotations

4. Isoform-level analysis report covering differential transcript usage (DTU) analysis between experimental conditions with statistical testing, alternative splicing event quantification with event-level significance, fusion transcript calls with breakpoint coordinates and supporting read evidence, and functional annotation of novel isoforms including coding potential and domain identification

5. Integrated multi-platform report (optional, for projects combining cDNA sequencing with other methods) correlating isoform expression with RNA modification status from parallel direct RNA sequencing, or providing cross-platform isoform validation comparing cDNA and PacBio RNA sequencing data for isoforms of interest