# Superb-seq from Salk reads Cas9 edits in single cells and finds 36 off-target sites in K562 cell line

- Published: 2026-11-16 18:37 UTC
- Section: Life & Health › Biotechnology
- Tags: Superb-seq, Cas9 off-target edits, single-cell RNA sequencing, Perturb-seq, T7 in situ transcription, K562 cells, genome editing, Sheriff analysis package
- URL: https://realtech.media/news/rt-e10152814304a47c9d4237d0_superb_seq_from/

**The method links each Cas9 edit to expression changes in the same cell. It was tested on 35,000 cells from three cell lines.**

![Superb-seq from Salk reads Cas9 edits in single cells and finds 36 off-target sites in K562 cell line](https://realtech.media/media/2026/11/b9b3a88e47f984de.jpg)
*techtimes.com*

Researchers at the Salk Institute describe Superb-seq, a method that captures Cas9 edits and their transcriptome effects in single cells, according to nature.com. It combines Cas9 editing, T7 in situ transcription and single-cell RNA sequencing. The authors ran it on 35,000 cells from three cell lines [1].

## Why guide capture falls short

Perturb-seq pairs Cas9 editing with single-cell RNA sequencing, but it records the guide RNA as a proxy for the edit rather than the edit itself. Guide detection is imperfect, and on-target and off-target activity varies. In some cell types and conditions up to 50% of Perturb-seq guides show no effect on the intended gene, which dampens estimates of regulatory effects [1]. Current off-target checks for therapy rely on computational prediction and bulk sequencing, and they do not measure the effect of an off-target edit on gene expression [1].

Superb-seq performs edit capture in three steps. Live cells receive Cas9 ribonucleoprotein and a donor DNA carrying the T7 promoter, which is inserted at the break by end joining without homology. Fixed cells then undergo T7 in situ transcription, so each labeled edit produces T7 transcripts. These are read together with endogenous transcripts by combinatorial single-cell RNA sequencing. The companion analysis package is called Sheriff [1].

The authors screened 27 donor constructs with T7 and SP6 promoters. Seventeen of them gave no detectable labeling. Three designs (01, 02 and 03) were 5.7–6.4 times more efficient than the GUIDE-seq donor, and design 02 was selected [1].

## Off-target results in K562 cells

In 10,000 K562 cells, seven guide RNAs targeted four chromatin remodeler genes. Superb-seq detected 5,761 edited cells, all on-target sites and 36 off-target sites. It estimated downstream effects on gene expression better than Perturb-seq [1].

The authors compared it with four bulk off-target detection methods on three benchmarking guides. Those methods identified between 6 and 210 edit sites [1]. Nineteen off-target edits were associated with differential gene expression, and nine of those were cell type-specific [1].

The work was done on cell lines, so it is a research tool rather than a clinical test. According to the authors, it connects single-cell edits with changes in gene expression [1].

## Sources

1. [nature.com](https://www.nature.com/articles/s41592-026-03319-z)
2. [techtimes.com](https://www.techtimes.com/articles/330808/20261116/crispr-off-target-edits-salks-superb-seq-finds-36-unintended-sites-one-tied-myc-gene-changes.htm)

---

Source: RealTech, realtech.media — https://realtech.media/news/rt-e10152814304a47c9d4237d0_superb_seq_from/
