Single-Cell Full-Length Transcriptome Sequencing (MAS-seq) — Isoform-Level Resolution at Single-Cell Scale

Most single-cell transcriptome studies today rely on 10x Genomics Chromium followed by Illumina short-read sequencing. This workflow captures gene expression counts with high throughput — but it only reads the 3' or 5' ends of transcripts. The internal exon structure, alternative splicing patterns, and full-length isoform identities remain invisible. A gene appearing "upregulated" in your UMAP may actually express a completely different isoform than the one you assume — and you would never know.

The gap between gene-level and isoform-level resolution is not incremental. It determines which biological questions your data can answer. Alternative last exons — found in over 50% of human genes — can switch a receptor from membrane-bound to secreted. Retained introns can introduce premature stop codons that trigger nonsense-mediated decay. Fusion transcripts arising from genomic rearrangements can produce chimeric proteins with oncogenic activity. Standard short-read scRNA-seq detects none of these events at single-cell resolution.

Our single-cell full-length transcriptome sequencing service closes this gap. By integrating 10x Genomics single-cell capture with PacBio Kinnex (MAS-seq) long-read sequencing, we deliver per-cell, per-isoform resolution — enabling you to distinguish splice variants, detect fusion transcripts, and quantify isoform switching across cell types. This is not a replacement for gene-level scRNA-seq; it is the layer of biological resolution that gene counts alone cannot provide. For bulk-level transcriptome projects, our full-length transcript isoform sequencing (Iso-Seq) service provides a high-throughput alternative.

Why Short-Read Single-Cell RNA-seq Cannot Resolve Isoform-Level Biology

Most single-cell transcriptome studies today rely on 10x Genomics Chromium followed by Illumina short-read sequencing. This workflow captures gene expression counts with high throughput — but it only reads the 3' or 5' ends of transcripts. The internal exon structure, alternative splicing patterns, and full-length isoform identities remain invisible. A gene appearing "upregulated" in your UMAP may actually express a completely different isoform than the one you assume — and you would never know.

The gap between gene-level and isoform-level resolution is not incremental. It determines which biological questions your data can answer. Alternative last exons — found in over 50% of human genes — can switch a receptor from membrane-bound to secreted. Retained introns can introduce premature stop codons that trigger nonsense-mediated decay. Fusion transcripts arising from genomic rearrangements can produce chimeric proteins with oncogenic activity. Standard short-read scRNA-seq detects none of these events at single-cell resolution.

Our single-cell full-length transcriptome sequencing service closes this gap. By integrating 10x Genomics single-cell capture with PacBio Kinnex (MAS-seq) long-read sequencing, we deliver per-cell, per-isoform resolution — enabling you to distinguish splice variants, detect fusion transcripts, and quantify isoform switching across cell types. This is not a replacement for gene-level scRNA-seq; it is the layer of biological resolution that gene counts alone cannot provide.

Technology Overview — The Complete Single-Cell Full-Length Transcriptome Workflow

Step 1: Single-Cell Isolation and cDNA Generation (10x Genomics Chromium)

For Option A projects, we begin with a single-cell suspension at ≥ 85% viability and ≥ 1 × 10⁶ cells. Cells are encapsulated into Gel Bead-in-Emulsion (GEM) droplets using the 10x Chromium X or Controller instrument. Each droplet contains a single cell, lysis reagents, and a gel bead carrying a unique 10x cell barcode and Unique Molecular Identifier (molecular barcode). Within each droplet, mRNA is reverse-transcribed into full-length cDNA, incorporating both the cell barcode and molecular barcode into every transcript molecule. After droplet breaking and cDNA amplification, the resulting library contains barcoded full-length cDNA from every captured cell.

For Option B projects, we receive your pre-amplified 10x cDNA and perform QC (Bioanalyzer concentration and size distribution) before proceeding directly to library preparation.

Step 2: Kinnex cDNA Concatenation and SMRTbell Library Preparation

Amplified 10x cDNA is processed through the PacBio Kinnex single-cell RNA kit. The workflow enzymatically ligates individual barcoded cDNA molecules into long concatemers, each containing cDNA from 8–16 original transcript molecules. These concatemers are then converted into SMRTbell libraries — hairpin adapters are ligated to both ends, creating a circular template for PacBio sequencing. Size selection using the SageELF or BluePippin system removes short fragments, enriching for full-length concatemers.

Step 3: PacBio Revio HiFi Sequencing

SMRTbell libraries are loaded onto SMRT Cell 8M and sequenced on the PacBio Revio system. During sequencing, each circular template is read multiple times by the polymerase, generating subreads that are computationally combined into a single high-accuracy HiFi read (≥ Q30, 99.9% base accuracy). Each SMRT Cell 8M produces approximately 90 Gb of HiFi data, with the concatenation strategy yielding roughly 16× more individual transcript reads than would be obtained from sequencing un-concatenated cDNA.

Step 4: Single-Cell Isoform Bioinformatics Analysis

The raw HiFi reads enter our specialized bioinformatics pipeline. Concatenated reads are first computationally split into individual cDNA molecules. Each cDNA read is then processed to recover its 10x cell barcode and molecular barcode, mapped to the reference genome/transcriptome, and assigned to a transcript isoform. molecular barcode deduplication corrects for PCR amplification bias, producing a cell-by-isoform expression matrix analogous to the cell-by-gene matrix from short-read scRNA-seq — but at isoform resolution. Downstream analyses include differential isoform expression across cell clusters, alternative splicing event detection, and fusion transcript identification.

Complete single-cell full-length transcriptome workflow: from 10x single-cell capture through Kinnex concatenation, PacBio Revio sequencing, and single-cell isoform bioinformatics analysis.

Long-Read vs Short-Read — Choosing the Right Single-Cell Transcriptome Approach

Selecting the right technology for your single-cell transcriptome study depends on whether you need gene-level counts or isoform-level resolution. The table below compares the three main approaches: standard short-read scRNA-seq, traditional scIso-Seq, and MAS-seq/Kinnex (our service).

How to choose:

• If your question requires only gene expression counts and cluster-based cell type identification, standard 10x + Illumina scRNA-seq remains sufficient.
• If you need full-length isoform information — to detect splice variants, isoform switching, or fusion transcripts — MAS-seq/Kinnex provides the isoform-level resolution you need at a cost that no longer makes it prohibitive.
• If you already have 10x cDNA from a previous experiment, you can add isoform-level resolution retrospectively by submitting it to our Kinnex pipeline.
• For researchers exploring alternative long-read platforms, we also offer Nanopore full-length cDNA sequencing as a complementary approach.

For bulk-level full-length transcriptome projects without single-cell resolution, our full-length transcript isoform sequencing (Iso-Seq) service provides a high-throughput alternative.

Case Study: Single-Cell and Spatial Alternative Splicing Analysis with Long-Read Sequencing

Fu Y, Kim H, Roy S, Huang S, Adams JI, Grimes SM, Lau BT, Sathe A, Ji HP, Zhang NR. Single cell and spatial alternative splicing analysis with Nanopore long read sequencing. Nature Communications. 2025;16:6654.

Background

Alternative splicing generates multiple transcript isoforms from a single gene, and isoform choices often differ between cell types, disease states, and tissue regions. However, measuring isoform-level expression at single-cell resolution has remained technically challenging. Short-read scRNA-seq captures gene counts but loses splicing information. Long-read sequencing can resolve full-length isoforms, but until recently, it suffered from high error rates that complicated cell barcode recovery and isoform quantification.

This study developed Longcell — a statistical and computational framework that enables accurate isoform quantification from single-cell and spatial barcode-labeled Nanopore long-read data. The authors applied Longcell to multiple datasets, including single-cell Jurkat T-cells, colorectal cancer liver metastasis samples, and spatial transcriptomics data, demonstrating its ability to quantify isoform expression, detect splicing factor regulatory targets, and resolve spatial isoform patterns.

Methods

The Longcell workflow integrates several computational innovations. First, it recovers cell barcodes and molecular barcodes from Nanopore long reads using an error-tolerant matching algorithm, achieving 50–70% read assignment to valid cell barcodes. Second, it corrects Nanopore sequencing errors that would otherwise misalign reads or truncate transcript annotations. Third, it applies a beta-binomial model to decompose total isoform variation into intracellular heterogeneity (multiple isoforms expressed within the same cell) and intercellular heterogeneity (different isoforms in different cell types).

To validate Longcell's ability to identify splicing factor targets, the authors performed CRISPR knockout of nine splicing factors in Jurkat T-cells, followed by single-cell Nanopore sequencing. Differential splicing analysis compared knockout and control cells, and key targets were validated by targeted amplicon sequencing.

Results

Longcell revealed that for highly expressed genes, intracellular isoform heterogeneity is the norm rather than the exception — most cells co-express multiple transcript isoforms from the same gene simultaneously. This challenges the assumption that individual cells express only one dominant isoform per gene.

Using colorectal cancer liver metastasis data, Longcell quantified isoform expression across four major cell types. The MYL6 gene provided a striking example: epithelial cells co-expressed both MYL6-218 and MYL6-207 isoforms, while myeloid cells exclusively expressed MYL6-207. This cell-type-specific isoform switching would be completely invisible in gene-level scRNA-seq data.

In the splicing factor perturbation experiment, Longcell identified significant differential splicing events for five genes (unstimulated Jurkat) and six genes (stimulated Jurkat) across nine CRISPR-targeted splicing factors. Remarkably, PCBP2 knockout promoted inclusion of exons 3 and 4 in the DGUOK gene — a previously unknown regulatory relationship that was subsequently validated by targeted sequencing. The study also confirmed the known regulation of PTPRC (CD45) alternative splicing by HNPRLL, providing internal validation of the method.

Longcell workflow for single-cell isoform quantification: cell barcode recovery from long reads, isoform-level expression matrix, and splicing factor perturbation analysis validated by targeted sequencing (adapted from Fu et al. 2025, Nature Communications).

Conclusions

This study demonstrates that single-cell long-read sequencing, combined with appropriate statistical modeling, can deliver isoform-level quantification that reveals regulatory relationships invisible to short-read methods. The discovery that intracellular isoform heterogeneity is widespread — even for highly expressed genes — has important implications for understanding the functional complexity of single-cell transcriptomes. For researchers designing single-cell studies where isoform-level information matters — including cancer biology, immunology, and neuroscience — this work provides both a methodological framework and biological evidence supporting the value of long-read single-cell isoform analysis.

Our service operationalizes these insights through an integrated 10x + PacBio Kinnex workflow. While the Longcell study used Nanopore sequencing, our PacBio Revio HiFi platform offers complementary advantages in base accuracy (≥ Q30), which directly benefits isoform assignment and fusion transcript detection. For projects requiring the spatial-temporal dimension of full-length transcriptome profiling, we offer integrated spatial+single-cell analysis workflows.

Demo Data Showcase

1. Isoform-Level UMAP Visualization

2. Splice Junction Heatmap Across Cell Types

3. Fusion Transcript Circos Plot