Cas9 can cleave DNA in both blunt and staggered configurations, resulting in distinct editing outcomes, but what dictates the type of Cas9 incisions is largely unknown. In this study, we developed BreakTag, a versatile method for profiling Cas9-induced DNA double-strand breaks (DSBs) and identifying the determinants of Cas9 incisions. Overall, we assessed cleavage by SpCas9 at more than 150,000 endogenous on-target and off-target sites targeted by approximately 3,500 single guide RNAs. We found that approximately 35% of SpCas9 DSBs are staggered, and the type of incision is influenced by DNA:gRNA complementarity and the use of engineered Cas9 variants. A machine learning model shows that Cas9 incision is dependent on the protospacer sequence and that human genetic variation impacts the configuration of Cas9 cuts and the DSB repair outcome. Matched datasets of Cas9 and engineered variant incisions with repair outcomes show that Cas9-mediated staggered breaks are linked with precise, templated and predictable single-nucleotide insertions, demonstrating that a scission-based gRNA design can be used to correct clinically relevant pathogenic single-nucleotide deletions.

Cas9 可以以平头和交错结构切割 DNA，从而产生不同的编辑结果，但决定 Cas9 切口类型的因素很大程度上未知。在这项研究中，我们开发了 BreakTag，这是一种通用方法，用于分析 Cas9 诱导的 DNA 双链断裂 (DSB) 并识别 Cas9 切口的决定因素。总体而言，我们评估了 SpCas9 对约 3,500 个单向导 RNA 靶向的超过 150,000 个内源性在靶和脱靶位点的切割。我们发现大约 35% 的 SpCas9 DSB 是交错的，并且切口的类型受到 DNA:gRNA 互补性和工程 Cas9 变体的使用的影响。机器学习模型表明 Cas9 切口取决于原型间隔序列，并且人类遗传变异会影响 Cas9 切口的配置和 DSB 修复结果。 Cas9 和工程变异切口与修复结果的匹配数据集表明，Cas9 介导的交错断裂与精确、模板化和可预测的单核苷酸插入相关，这表明基于断裂的 gRNA 设计可用于纠正临床相关的致病性单核苷酸删除。