Demand for high-resolution transcriptome data has risen sharply as questions shift from which genes are expressed to which isoforms, where they are expressed, and when they are expressed. Long-read platforms, such as those developed by Oxford Nanopore Technologies and Pacific Biosciences (PacBio), now deliver individual cDNA molecules of several kilobases, promising an unfiltered view of splicing, fusion events, and allele-specific expression.
In parallel, refinements of short-read “full-length” protocols, including VASA-seq, offer uniform coverage at Illumina-class accuracy and a scale suitable for entire cell atlases.
This blog outlines the strengths and limitations of each strategy, guiding the selection of the most suitable method for a specific biological question. You can always contact us with a specific question or if you would like to discuss your situation with our experts.
What is long-read sequencing?
Long-read sequencing platforms generate DNA or RNA reads spanning tens of thousands of base pairs, compared to the few hundred bases typically generated by Illumina platforms. Companies like Oxford Nanopore (ONT) and PacBio have made this technology increasingly accessible, with ONT offering portable devices that can run in any lab and PacBio providing high-throughput systems for large-scale projects. The real value lies in resolving complex genomic regions that confound short-read assembly, such as repetitive sequences, structural variants, and full-length transcript isoforms. This makes long-read sequencing particularly powerful for applications like cancer genomics, where structural rearrangements drive disease, or developmental biology, where alternative splicing creates functional diversity from a limited number of genes.
Why choose long-read sequencing?
Short-read methods break each transcript into tiny fragments that must be stitched back together in silico, an approach that can obscure splice variants, structural rearrangements, and subtle allelic differences. Long-read platforms capture entire RNA molecules in one pass, allowing researchers to inspect the full transcript directly instead of inferring it from fragments. Falling sequencing costs and simpler library prep now make this once-specialized technology practical for everyday projects.
Benefits of long-read sequencing platforms
Systems from Oxford Nanopore and PacBio deliver single-molecule, isoform-level views of the transcriptome. They resolve closely related variants, uncover rare or complex splicing events such as exon skipping and intron retention, and supply direct experimental evidence for previously unannotated isoforms, strengthening genome annotations without relying on error-prone computational assembly.
Limitations of long-read sequencing in single-cell research
Long-read sequencing solves many annotation problems, yet the current generation of platforms still faces several constraints when applied to large single-cell data sets.
-
High per-cell cost and limited scalability
Library preparation and flow-cell expenses remain several times higher than short-read workflows, especially when a project requires tens of thousands of cells. -
Higher raw error rates
Even with improved chemistries, single-pass accuracies of 97–99 % lag behind the 99.9 % typical of Illumina sequencing. Error-correction can recover base accuracy but adds compute time and may trim read depth. - Lower throughput for high-cell-number studies
A single PromethION or Revio run usually delivers data for 5 000–20 000 cells. Large atlases, screens, or time-course experiments therefore demand multiple flow-cells and a larger budget. - Extra experimental steps
Size selection, additional clean-ups, and long amplification cycles lengthen protocols and raise the risk of sample loss, particularly when dealing with rare or fragile cell types. - Complex data analysis and integration
Long-read alignment, isoform collapsing, and polishing require bespoke tools and considerable compute resources. Integrating long-read output with established short-read pipelines (for example, Seurat or Scanpy) still involves custom workflows.
VASA-seq: a total-RNA alternative to long-read sequencing
While long-read sequencing provides a comprehensive view of each transcript in a single uninterrupted sweep, its high cost and limited throughput can hinder discovery studies that require thousands of cells. VASA-seq (Vast Transcriptome Analysis of Single cells by dA-tailing) can close this gap by combining full-length coverage with the accuracy and scale of Illumina sequencing.
VASA-seq takes a simple but effective detour around the need for a natural poly(A) tail. After cell lysis, every RNA fragment receives a short artificial dA tail. A standard oligo-dT primer then initiates reverse transcription, which means:
- Poly(A) independence
Histone mRNAs, many long non-coding RNAs, snoRNAs, and partially processed pre-mRNAs all enter the library. Studies that focus on regulatory or rapidly turned-over transcripts benefit in particular. - No template-switch step
First-strand synthesis proceeds directly, avoiding the bias that often under-represents 5′ ends in conventional full-length protocols. Coverage is therefore more even across each gene body, improving quantification of untranslated regions and splice junctions. - Compatibility with standard sequencers
cDNA is fragmented and sequenced on Illumina instruments, so each base is called with an error rate of approximately 0.1%. This accuracy supports the detection of single-nucleotide variants and facilitates straightforward integration with existing short-read pipelines. - Flexible scale
The chemistry works equally well in microtiter plates as in droplets. A pilot on a few hundred rare cells or a census of tens of thousands can be run without specialised hardware.
When VASA-seq outperforms long-read workflows
- Projects that ask “which exons are present?” rather than “which exon order is used?”
- Experiments centred on non-coding or rapidly degraded RNAs that lack poly(A) tails.
- Studies that need uniform coverage for transcript-velocity analysis or precise 5′ UTR quantification.
- Laboratories that prefer established Illumina-based analysis pipelines.
Selecting the optimal workflow
Long-read sequencing remains the method of choice when splice-variant architecture or fusion breakpoints are the primary read-outs. VASA-seq offers broader molecular reach and higher sensitivity at a fraction of the cost per cell. Many research groups now adopt a hybrid strategy: first use VASA-seq to understand single-cell exon usage, then use a targeted long-read experiment to catalogue full isoforms. The optimal balance depends on the biological hypothesis, available budget, and desired cell count.
If you would like guidance on choosing or combining these methods, our team is ready to discuss the specifics of your samples and research goals.
Find out more about VASA-seq in our whitepaper
