Superb-seq links Cas9 edits to gene expression in single cells: 36 off-target sites in K562 cell line
The method combines Cas9 editing, T7 in situ transcription and single-cell RNA sequencing in cell lines, reading edits directly rather than inferring them from guide RNAs.

A study available via doi.org describes Superb-seq, a method that records Cas9 edits and the transcriptome in the same single cells nature.com
doi.orgnature.comOpen source ↗. The authors ran it on 35,000 cells from three cell lines in two experiments. In one of them, seven guide RNAs targeted four chromatin remodeler genes in 10,000 K562 cells. Superb-seq detected 5,761 edited cells, all on-target sites and 36 off-target sites.
How edits are captured
Perturb-seq methods use guide RNAs as a proxy for edits, which leaves ambiguity between a detected guide and an actual edit event. The authors cite reports that in some conditions up to 50% of Perturb-seq guides show no effect on the intended target gene. Superb-seq instead performs "edit capture" in three steps. Live cells are edited with a Cas9 ribonucleoprotein and a donor DNA carrying a T7 promoter, which is inserted at the break by homology-free end joining. Fixed cells then undergo T7 in situ transcription, producing edit-reporting T7 transcripts in each cell. Finally, T7 and endogenous transcripts are read together by combinatorial single-cell RNA sequencing.
To choose the donor, the authors screened 27 constructs encoding T7 and SP6 promoters. Seventeen of them produced no detectable labeling. Three designs combining elements of the GUIDE-seq donor and optimized T7 promoters showed 5.7–6.4-fold higher efficiency than the GUIDE-seq donor, and design 02 was selected. Data are analyzed with the companion package Sheriff.
Off-target sites and expression
According to the authors, Superb-seq estimated downstream effects of perturbations on gene expression better than Perturb-seq nature.com
doi.orgnature.comOpen source ↗. They also compared it with four bulk off-target detection methods; across three benchmarking guides, the methods identified between 6 and 210 edit sites.
Nineteen off-target edits were associated with differential gene expression, and nine of these were cell type-specific. The authors note that current assessment of therapeutic off-targeting relies on computational prediction and bulk sequencing, which does not measure effects on gene expression. The results were obtained in cell lines.
Sources · 2
Found a mistake?Read also



