Salk Institute scientists have introduced Superb-seq, a new genome-profiling method that reads out both on-target and off-target CRISPR/Cas9 edits alongside changes in cellular gene activity at single-cell resolution. The peer-reviewed study, published in Nature Methods on November 16, 2026, describes a platform created to identify which individual cells carry specific genetic alterations and determine how those modifications influence the cell’s overall molecular landscape. By connecting precise physical edits directly to their downstream effects on gene expression, the method addresses a longstanding technical blind spot in genome engineering, where the cellular impacts of unintended edits have remained largely unobservable using standard technologies.
Genome-editing research accelerated in 2012 following the discovery that CRISPR/Cas9 could operate as molecular scissors to cut designated DNA sequences. In natural bacterial systems, CRISPR elements and the Cas9 protein serve as a defense mechanism against infection by using a guide RNA strand to target and cut viral genetic material. Researchers recognized that synthetic guide RNAs could be engineered to direct cutting to any chosen segment of genetic code, offering capabilities to delete genes for laboratory functional studies or replace mutated sequences to restore normal function. Although this approach fostered gene-editing therapies—including the first FDA-approved CRISPR treatment in 2023 for sickle cell disease and beta thalassemia—unwanted off-target cutting has presented a persistent technical obstacle for clinical and basic applications.
Existing edit-detection methods can pinpoint the genomic coordinates of off-target edits across a cell population, but they cannot identify which specific cells harbor those edits or assess how the edits influence cell behavior. Senior author Graham McVicker, an associate professor at the Salk Institute, indicated that frustration with this analytical limitation motivated the creation of Superb-seq. McVicker noted that while the scientific community recognized the challenge of CRISPR making unintended cuts, researchers lacked an effective tool to observe what those specific edits were and how they altered the cell, leaving those molecular consequences invisible under current experimental workflows.
The study also connects to the historical genetics legacy at the Salk Institute, an independent nonprofit research institution established in 1960 by Jonas Salk following his development of the polio vaccine. Salk was formerly home to Francis Crick, who earned a Nobel Prize as co-discoverer of the double-helical structure of DNA and later served as research professor and president of the Institute. McVicker’s laboratory sought to build upon this foundational research history by addressing contemporary questions regarding how genome editing can be rendered more precise for scientists in laboratory discovery and safer for patients receiving genetic therapies.
Mechanistically, Superb-seq functions by integrating Cas9 genome editing, a short edit-reporting DNA sequence, and single-cell RNA sequencing into an integrated protocol. This unified design captures both individual edits and their direct consequences on gene expression within each targeted cell. In doing so, the system detects alterations in gene expression driven by genome edits that remain undetectable using existing edit-detection approaches, providing an empirical bridge between physical sequence alterations and observed transcriptomic changes.
The platform offers critical analytical advantages over earlier technological platforms, including Perturb-seq. While Perturb-seq reads out cellular gene expression and delivered guide RNAs simultaneously, it does not detect the actual genetic edits that result from the cutting process. In contrast, Superb-seq measures both the on-target and off-target edits as well as the resulting changes in gene activity at single-cell resolution, establishing a more detailed assessment of how genomic interventions reshape the cell’s molecular landscape.
The Salk research team validated Superb-seq by testing it on 34,000 cells drawn from three distinct laboratory cell lines. In an evaluation conducted in 10,000 human leukemia cells grown in the laboratory, Superb-seq identified 43 edit sites—36 of which were off-target locations—across 5,761 edited cells. McVicker noted that while investigators who have worked on off-target effects previously possess an awareness of the problem, the sheer frequency of off-target cuts revealed by the platform exceeded expectations and may surprise many researchers who routinely employ CRISPR in the laboratory.
Co-first author Brad Balderson, a postdoctoral researcher in McVicker’s laboratory, explained that off-target edits can potentially disrupt critical genes that protect against cancer or other unhealthy cellular states. Balderson pointed out that it is necessary to not only identify which off-target edits occur, but also to ascertain which individual cells contain them and whether those changes disrupt activities that could produce disease. By enabling scientists to distinguish between likely benign modifications and consequential off-target edits, Superb-seq supplies an essential evaluative layer for assessing experimental and therapeutic interventions.
The research team outlined four primary applications for Superb-seq: basic scientific research, drug development, personalized medicine, and training machine learning algorithms. Co-first author Michael Lorenzini, a staff scientist in McVicker’s laboratory, observed that tens of thousands of published biological studies utilize CRISPR, suggesting that Superb-seq could help clarify the degree to which off-target edits have colored those collective findings. In therapeutic settings, where individual genetic variations can generate patient-specific off-target sites, the platform could be deployed during early preclinical stages to profile candidates and characterize both their off-target cutting profile and functional effects to identify optimal treatment options.
The authors also highlighted Superb-seq as a high-quality data generation platform for training artificial intelligence models that predict cellular responses to genetic perturbations. The Salk researchers emphasized that while recent computational efforts have focused heavily on increasing training data volume, the absence of single-cell edit outcomes in existing datasets constrains what machine learning models can learn. McVicker, Balderson, and Lorenzini have filed a patent for Superb-seq. The study, co-authored by Shanna Lavalle, Karthyayani Sajeev, and Aaron Ho of Salk, alongside Madelyn Light, Olalekan Usman, and Gavin Kurgan of Integrated DNA Technologies, received funding from the National Institutes of Health and multiple philanthropic foundations.
The development of Superb-seq reflects an ongoing evolution across more than a century of genetics research, which progressed from recognizing that biological inheritance is encoded in a four-letter molecular alphabet to uncovering the double-helical architecture of DNA and mapping the genetic roots of human pathology. When researchers adapted CRISPR from microbial systems into a programmable genome-editing apparatus, the technology raised prospects for permanently correcting severe monogenic conditions, including cystic fibrosis and Huntington’s disease. Yet as genomic tools shifted from merely reading biological sequences toward actively rewriting cellular instructions, the challenge of maintaining sequence fidelity became paramount. Unintended cuts elsewhere in the genome demonstrated that manipulating complex biological systems requires an exact accounting of every molecular modification introduced during an experiment or clinical intervention.
The underlying mechanism of CRISPR genome editing originated as a natural bacterial immune response designed to neutralize invading pathogens by cleaving viral genetic sequences. In natural bacterial environments, the Cas9 enzyme executes targeted double-stranded cuts guided by short RNA molecules transcribed from clustered regularly interspaced short palindromic repeat elements. While laboratory investigators repurposed this molecular machinery by engineering customized synthetic guide RNAs to target specific genes for removal or therapeutic replacement, Cas9 can still bind and cut unintended genomic loci that resemble the intended target. When these off-target events disrupt vital protective elements, such as tumor suppressor pathways guarding against cancer or other unhealthy cellular states, the consequences can compromise experimental validity or clinical safety.
This biological reality underscores the importance of evaluating genome editing within personalized medical frameworks. Because individual human genomes harbor distinct sequence variations, a guide RNA that performs cleanly in one individual might induce off-target cutting at novel sites in another patient with a slightly different genetic background. Senior author Graham McVicker highlighted that the safest therapeutic strategy will inherently differ between patients due to these subtle individual genetic discrepancies. By profiling candidate therapies across individual cells, Superb-seq allows researchers to identify both unintended edits and their cellular consequences on a patient-specific basis. This capability provides an empirical method to evaluate candidate gene therapies during early preclinical development, helping scientists screen out dangerous modifications before advancing experimental treatments toward patient administration.
The method also addresses a critical bottleneck in computational biology, where artificial intelligence architectures are trained to forecast how cells respond to genetic perturbations. Recent initiatives across the machine learning community have placed substantial emphasis on expanding the raw volume of training data. However, the Salk team pointed out that improving data quality is equally vital, because the absence of single-cell edit outcomes in existing reference libraries severely restricts what machine learning models can learn. By systematically pairing physical sequence alterations with single-cell transcriptomic readouts, Superb-seq provides the higher-fidelity training data required to refine artificial intelligence models of cellular function and perturbation biology.
Such foundational tool development reflects the institutional mission of the Salk Institute, founded in 1960 by Jonas Salk following his creation of the polio vaccine, to advance risk-taking research addressing widespread challenges in cancer, Alzheimer’s disease, and crop resilience. Looking to bridge discoveries from bench to bedside, McVicker, Brad Balderson, and Michael Lorenzini filed a patent for the method. The investigation was funded through National Institutes of Health grants HG011315, P30 AG068635, P01 AG073084-04, P30 CA014195, S10 OD023689, and S10 OD034268, with philanthropic backing from the Frederick B. Rentschler Developmental Chair, the George E Hewitt Foundation for Medical Research, the Howard and Maryam Newman Family Foundation, the Helmsley Charitable Trust, and the Chapman Foundation.
Subject of Research: Single-cell profiling of CRISPR genome editing and gene expression responses
Article Title: Can genome editing be safer and more effective?
Article References: Can genome editing be safer and more effective?. Original publication (Bibliographic details unavailable.)
Image Credits: AI Generated
DOI: Not provided
Keywords: Superb-seq, CRISPR-Cas9, genome editing, single-cell RNA sequencing, off-target edits, gene expression, Salk Institute, Graham McVicker, Brad Balderson, Michael Lorenzini, Perturb-seq, Nature Methods, cell lines, human leukemia cells, machine learning models, personalized medicine, drug development, gene regulation
News Source: Juliet Wilcox. (November 17, 2026). Superb-seq Detects On- and Off-Target CRISPR Edits at Single-Cell Resolution. Scienmag.



