Strategy for detecting off-target sites in genome-edited rice

Genome-editing using the CRISPR-Cas9 system can substantially accelerate crop breeding. Because off-target editing is the main problem with this system, a reliable method for comprehensively detecting off-target sites is required for the editing of food crop genomes. However, a method that accurately predicts off-target sites has not been established. In this study, we performed a SITE-Seq analysis to predict potential off-target sites. SITE-Seq is an unbiased method applicable for the in vitro detection of double-strand breaks (DSBs). To analyze SITE-Seq data, we developed a novel Galaxy system, which can perform simple and reproducible analyses without a command line operation. We conducted a SITE-Seq analysis of a rice genome modified by OsFH15 gRNA-Cas9, and identified 41 DSB sites in the annotated regions. Amplicon-sequencing revealed mutations at one off-target site in the genome-edited rice. The presence of an uncommon protospacer adjacent motif (NTG PAM) likely makes this off-target site difficult to identify using in silico methods. Of the six tested programs, only CRISPRdirect predicted this off-target site, but it also predicted 6,080 off-target sites in total. These results suggest the SITE-Seq method presented herein can efficiently predict off-target sites and is useful for assessing the safety of genome-edited food.