Skip to main content
CRISPR TechnologySuperb-seq· 6 min read· in Science

New Sequencing Tool Tracks Off-Target CRISPR Gene Edits Across Individual Cells

Superb-seq combines Cas9 edit detection with single-cell RNA sequencing to reveal how unintended DNA cuts alter gene expression. The platform identified dozens of off-target mutations missed by traditional bulk sequencing methods.

By Sofia Matos

Share this story

Listen

A new sequencing platform developed at the Salk Institute allows researchers to track exactly where CRISPR-Cas9 cuts DNA and how those cuts alter cellular behavior. The tool, called Superb-seq, captures both the physical genetic edit and the resulting gene expression changes within individual cells.[1][6]

While CRISPR-Cas9 has transformed genetic engineering, unintended off-target edits remain a critical safety hurdle for clinical applications. Identifying these stray cuts has traditionally required researchers to choose between mapping the locations of the edits or measuring their functional impact on the cell.[1][4]

“A longstanding challenge in functional genomics and therapeutic genome engineering has been the inability to directly evaluate CRISPR-Cas9 editing outcomes alongside their molecular consequences within the same single cells,” notes Bioengineer.org. Existing tools force a compromise that limits therapeutic development.[1]

Bulk sequencing methods can locate off-target cuts across a population of cells, but they lose the individual cellular context. Conversely, single-cell RNA sequencing tracks gene expression but struggles to confirm whether a specific cell actually received the intended DNA edit.[1][6]

Superb-seq bridges this gap by integrating Cas9 editing with T7 in situ transcription and combinatorial single-cell RNA sequencing. The methodology was developed by a research team led by Michael H. Lorenzini, Brad Balderson, and Graham McVicker.[1][6]

Tagging the DNA breaks

The foundation of the Superb-seq workflow relies on a modular, homology-free tagging system that marks Cas9-generated double-strand breaks. Researchers introduce a synthetic double-stranded donor DNA that encodes a bacteriophage T7 promoter sequence into the live cells.[1][4]

When the Cas9 enzyme cuts the genome, the cell’s natural repair pathways insert this synthetic T7 promoter directly into the break site. This insertion acts as a permanent barcode, labeling both intended on-target edits and unintended off-target cleavages.[1][6]

The platform inserts a synthetic T7 promoter at the site of a Cas9 cut, acting as a permanent barcode.

After the cells are fixed, the team uses T7 RNA polymerase to generate transcripts from these inserted promoters. These edit-marking transcripts are then sequenced alongside the cell’s natural messenger RNA using an off-the-shelf combinatorial indexing chemistry.[1][4]

To interpret the dual sequencing libraries, the researchers built an open-source analytical software suite called Sheriff. The software inspects the sequencing reads to detect the distinctive synthetic T7 barcodes, separating authentic edit markers from background cellular noise.[1][4]

Mapping off-target frequencies

To validate the platform, the Salk team tested Superb-seq on 10,000 human leukemia cells from the K562 line. They deployed seven distinct guide RNAs designed to target four specific chromatin remodeler genes, including ARID1A and SMARCA4.[1][6]

The targeted genes encode subunits of the BAF and NuRD chromatin remodeling complexes, which play essential roles in regulating DNA architecture. Disrupting these specific genes allowed the researchers to observe broad, cascading changes across the cellular transcriptome.[1][6]

The Sheriff software successfully identified 5,761 edited cells and recovered all seven of the intended on-target genomic cleavage sites. However, the high-resolution scan also revealed the extent of the unintended genetic modifications caused by the guides.[1][6]

The platform discovered 36 distinct off-target cleavage sites scattered across 18 different chromosomes in 2,098 of the cells. The frequency of these unintended cuts varied widely, appearing in anywhere from 0.05 percent to 18 percent of the edited population.[1][4][6]

Off-target edit frequencies varied widely across the tested cell populations.

“Cas9 off-targeting is pervasive due to a combination of rare and common edit events, occurs primarily within introns of off-target genes, and can exert widespread effects on gene expression,” the researchers detailed in their 2026 bioRxiv preprint. This finding underscores the hidden risks of seemingly precise gene therapies.[6]

Disrupting downstream pathways

The sequencing data revealed that more than 80 percent of the unintended editing events occurred within intronic sequences. While these non-coding regions do not directly produce proteins, they often contain vital regulatory elements that control other genes.[1][6]

One notable off-target cut landed in the first intron of the USP9X gene, a region responsible for regulating protein abundance. This single unintended edit disrupted the expression of USP9X and triggered a cascade that altered more than 150 downstream genes.[1][4][6]

Among the affected downstream targets, the MYC oncogene was significantly repressed by the intronic disruption. This specific pathway alteration demonstrates how off-target edits in non-coding regions can drastically rewire cellular functions and influence cancer-related gene networks.[1][6]

Beyond physical mapping, Superb-seq provides the statistical power to associate edit allele dosages with quantitative shifts in gene expression. The team used a linear mixed-effects model to confirm dose-dependent transcriptional downregulation for the on-target edits in ARID1A and CHD3.[1][6]

The statistical model effectively separated the direct consequences of the Cas9 cuts from natural cellular variations. By quantifying the exact dosage of the edit alleles, the software mapped a direct correlation between the genetic disruption and the resulting transcriptomic shift.[1][6]

Outperforming bulk sequencing

The researchers benchmarked Superb-seq against established single-cell perturbation screens, including a large-scale Perturb-seq study of 2.5 million K562 cells. Superb-seq demonstrated significantly superior sensitivity in capturing the expected downstream pathway responses determined by prior chromatin profiling.[1][6]

This heightened sensitivity stems from the platform’s ability to exclude unedited cells from the analysis. Standard guide-capture screens often include guide-positive cells that never actually sustained a DNA cut, which dilutes the measured effect sizes.[1]

The team further evaluated the tool against four established bulk off-target detection platforms, including GUIDE-seq and DISCOVER-seq. They utilized a 16,000-cell benchmark library and three standardized guide RNAs targeting specific sites in human cell lines.[1][6]

Across the benchmark library, Superb-seq captured 210 distinct edit sites for a standard guide targeting the VEGFA gene. This total included 74 off-target cleavage sites that the four bulk profiling methods had completely failed to detect.[1][6]

Superb-seq identified dozens of off-target cleavage sites that standard bulk profiling methods missed.

While Superb-seq excelled at finding frequent off-target cuts, its sensitivity varied based on the overall indel rate. The platform detected up to 80 percent of frequent edit sites with indel rates above ten percent, though detection dropped for rarer mutations below that threshold.[1]

Accessible therapeutic screening

Despite its high resolution, the Superb-seq protocol is designed to be highly accessible for standard research laboratories. The workflow relies entirely on off-the-shelf commercial reagents and standard equipment, avoiding the need for complex viral vectors.[1][6]

Because the method preserves the cell’s complete transcriptomic identity, researchers can mix different cell populations during the experiment. The Sheriff software successfully de-multiplexes these mixed populations computationally, bypassing the need for physical cell sorting prior to sequencing.[1][6]

This computational sorting capability allows scientists to test CRISPR therapies directly on complex, mixed tissues like primary patient samples. Researchers can observe exactly which cell types sustain off-target edits and how those specific cells react to the damage.[1][4]

The analytical pipeline is already being integrated with commercial single-cell sequencing platforms, including the Evercode system from Parse Biosciences. This broad compatibility ensures that laboratories can adopt the Superb-seq workflow without investing in specialized, proprietary hardware.[4]

The technology provides a rigorous new standard for vetting guide RNAs before they enter human clinical trials. By identifying hidden off-target effects early in the development pipeline, scientists can design safer, more precise genetic therapies.[1][2][3]

The technology provides a rigorous new standard for vetting guide RNAs before they enter human clinical trials.

The Salk Institute team envisions Superb-seq becoming a foundational tool for functional genomics research. Linking the physical reality of a DNA cut to the functional reality of gene expression creates a comprehensive map of CRISPR’s true impact.[1][5][6]

Key points

  • Superb-seq integrates Cas9 editing with T7 in situ transcription to map DNA cuts and their transcriptomic effects within the same cell.
  • The platform detected 36 off-target cleavage sites in a 10,000-cell test, revealing that most unintended cuts occur in non-coding intronic regions.
  • An off-target edit in the USP9X gene disrupted its expression and altered more than 150 downstream genes, including repressing the MYC oncogene.
  • The tool identified 74 off-target sites missed by standard bulk profiling methods, offering a more rigorous safety screen for clinical guide RNAs.

What we don’t know

  • How Superb-seq performs when tracking edits in highly complex, solid tissue samples outside of standard cell lines.
  • Whether the platform's sensitivity for detecting rare mutations below a ten percent indel rate can be improved in future iterations.
  • The long-term clinical impact of the specific non-coding off-target edits identified during the platform's initial benchmarking tests.
Functional Genomics Researchers 40%Therapeutic Developers 40%Computational Biologists 20%
Functional Genomics Researchers
Value the ability to map regulatory networks and understand how non-coding edits affect gene expression.
Therapeutic Developers
Focus on the safety profile of CRISPR therapies and the need to rigorously vet guide RNAs for off-target effects before clinical trials.
Computational Biologists
Emphasize the importance of open-source software like Sheriff to process complex, multi-modal single-cell datasets.

Perspectives this story doesn't cover

  • Clinical Trial Regulators
  • Patients Awaiting Gene Therapies

Sources

Source coverage

6 outlets

3 viewpoints surfaced

Functional Genomics Researchers 40%Therapeutic Developers 40%Computational Biologists 20%
  1. [1]Bioengineer.orgTherapeutic Developers

    Superb-seq Couples Single-Cell Cas9 Editing With Transcriptome Profiling to Uncover Off-Target Gene Regulation

    Read on Bioengineer.org →
  2. [2]Tom's Hardware ItaliaTherapeutic Developers

    Correggere il DNA può provocare errori, ora sappiamo dove cercarli

    Read on Tom's Hardware Italia →
  3. [3]WisevoterTherapeutic Developers

    Superb-seq Maps CRISPR Edits in Single Cells

    Read on Wisevoter →
  4. [4]Parse BiosciencesComputational Biologists

    Superb-seq Uncovers Hidden Off-Target Edits with CRISPR + Single-Cell Precision

    Read on Parse Biosciences →
  5. [5]Salk InstituteFunctional Genomics Researchers

    Keck Foundation funds three innovative early-career Salk projects

    Read on Salk Institute →
  6. [6]bioRxivFunctional Genomics Researchers

    Joint single-cell profiling of Cas9 edits and transcriptomes reveals widespread off-target events and effects on gene expression

    Read on bioRxiv →
Listen

Comments

Stay informed

Every angle. Every day.

Get Science stories with full source coverage and perspective breakdowns, free every day.