Back

Single-Cell and Tissue-Specific CRISPR Editing Analyses Unveil New Insights to Off-Targets and Translocations

Madsen, A.; Selfjord, N.; Martinez-Lage Garcia, M.; Loyd, A.-L.; Kurgan, G.; Stahlberg, M.; Lindgren, J.; Liz Touza, J.; Wigge, L.; Firth, M.; Nordstrom, K.; Collin, J.; Jachimowicz, D.; Schiffthaler, B.; Dillmann, I.; Antoniou, P.; Emmanouilidi, A.; Hellsten, J.; Forsstrom, J.; Magnell, K.; Jacobi, A.; Behlke, M.; Porritt, M.; Madeyski-Bengtson, K.; Maresca, M.; Akcakaya, P.

2026-04-15 genomics
10.64898/2026.04.13.718115 bioRxiv
Show abstract

CRISPR-Cas9 holds promise for treating genetic disease, but rare off-target mutations and structural variants remain as key safety concerns, especially at scales relevant to therapy. We established workflows to resolve Cas9 off-target activity in vitro at single-cell resolution and in vivo across different tissues. Using clonally expanded electroporated mouse embryos and embryonic stem cells, we reveal that individual cells exhibit unique off-target and translocation profiles, including events missed in bulk analyses. Integrating single-cell editing with chromatin accessibility, transcription, and DNA methylation measurements suggested that sequence-independent features modulate Cas9 access and cleavage, with preferential editing in regions characterized by open chromatin and lower methylation. In Cas9-inducible mouse models, editing analyses revealed organ-distinct off-target spectra, DNA repair pathway usage, indel patterns, and markedly varying translocation propensity between tissues. These findings demonstrate that off-target activity is heterogeneous across cells and context-dependent across organs, motivating sensitive single-cell analyses and organ-specific evaluation in preclinical development to more accurately assess risk and improve the safety of CRISPR-based genomic medicines.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.