bioRxivMolecular BiologyPreprint
Prime editing primarily induces undesired outcomes in mice
Tomomi AidaiD, Jonathan J. WildeiD, Lixin YangiD, Yuanyuan Hou, Mengqi LiiD, Dongdong Xu, Jianbang LiniD, Peimin Qi, Zhonghua LuiD, Guoping FengiD
Abstract summary
Claude Opus 5.5 · xhigh effort · 18 Oct 2026 · from the abstract · CC BY-NC-ND 4.0
In mouse zygotes, prime editing installed 10 modifications across 6 genes, but undesired products such as large deletions outnumbered intended edits, traced to PE3 double-nicking.
Prime editing is presented as a promising approach because of its proposed flexibility and its ability to avoid unwanted indels, and this work tests it in mammalian zygotes. Working in mouse embryos, the authors find that the dominant products are not the intended edits, and tie that to the second nick used by the PE3 strategy.
Key results
ShownPrime editing is functional in mammalian zygotes, installing substitutions, deletions and insertions when delivered with chemically modified guide RNAs.
10 targeted modifications across 6 genes were introduced in mouse embryos; the abstract calls the editing highly efficient but reports no efficiency percentages here. · Abstract
SuggestedUnintended products, including large deletions, dominated the outcome at every site tested, outnumbering clean intended edits.
Reported as occurring more frequently than pure intended edits across all edits and genes examined; no frequencies, allele counts or genotyping method are given in this text. · Abstract
SuggestedThe second nick of the PE3 strategy is responsible, but it cannot simply be dropped: removing it mostly abolishes editing, while sequential nicking with PE3b removes the unwanted products.
PE3b is reported to eliminate undesired outcomes but to be usable only for a fraction of the edits attempted; no numbers accompany these comparisons in this text. · Abstract
Methods, data and reproducibility
Prime editing components with chemically modified guide RNAs were delivered to mouse zygotes to install a panel of substitutions, deletions and insertions, and the resulting alleles were examined for intended versus unintended products while comparing the PE3 double-nicking strategy, nick omission and sequential PE3b nicking.
- Mouse zygotes/embryos edited with prime editing and genotyped for intended and unintended alleles · mouse · n = 10 targeted modifications across 6 genes
- Methods & tools
- prime editing (PE), PE3, PE3b, chemically modified guide RNAs
Limitations (4)
- The authors note that PE3b, the variant that removed unintended products, applies only to a fraction of possible edits. · stated by the authors
- The authors conclude that improved prime editing variants are needed before the approach is suitable for general use. · stated by the authors
- No frequencies, allele counts or genotyping assay are reported in the available text, so the rate and size spectrum of the unintended deletions cannot be judged. · noted in this summary
- Findings come from 10 edits at 6 loci in mouse zygotes; whether the same outcome balance holds at other loci, in other cell types or in other species is untested here. · noted in this summary
Read this if you plan prime editing in embryos or animal models and need to know whether PE3 yields large deletions rather than the intended edit.
- Citations
- 43
Abstract
Genome editing has transformed biomedical science, but is still unpredictable and often induces undesired outcomes. Prime editing (PE) is a promising new approach due to its proposed flexibility and ability to avoid unwanted indels. Here, we show highly efficient PE-mediated genome editing in mammalian zygotes. Utilizing chemically modified guideRNAs, PE efficiently introduced 10 targeted modifications including substitutions, deletions, and insertions across 6 genes in mouse embryos. However, we unexpectedly observed a high frequency of undesired outcomes such as large deletions and found that these occurred more often than pure intended edits across all of the edits/genes. We show that undesired outcomes result from the double-nicking PE3 strategy, but that omission of the second nick largely ablates PE function. However, sequential double-nicking with PE3b, which is only applicable to a fraction of edits, eliminated undesired outcomes. Overall, our findings demonstrate the promising potential of PE for predictable, flexible, and highly efficient in vivo genome editing, but highlight the need for improved variations of PE before it is ready for widespread use.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Optimized nickase- and nuclease-based prime editing in human and mouse cells
Fatwa Adikusuma, Caleb Lushington, … Paul Quinton Thomas·Nucleic Acids ResearchPeer-reviewed·Sep 16, 2021
- Prime editing enables precise genome editing in mouse liver and retina
H. Jang, J. H. Shin, … H. H. Kim·bioRxiv·Jan 8, 2021
- Improved nuclease-based prime editing by DNA repair modulation and pegRNA engineering
P. Antoniou, L. Dacquay, … M. Peterka·bioRxiv·Feb 1, 2024
- Error-free and efficient prime editing in absence of maternal Polθ
François Kroll, Malo Serafini, … Filippo Del Bene·bioRxiv·May 13, 2026
- No observable guide-RNA-independent off-target mutation induced by prime editor
- Improved prime editors enable pathogenic allele correction and cancer modelling in adult mice
P. Liu, S. Liang, … W. Xue·Nature CommunicationsPeer-reviewed·Apr 9, 2021