Whole genome sequencing reveals rare off-target mutations and considerable inherent genetic or/and somaclonal variations in CRISPR/Cas9-edited cotton plants

The CRISPR/Cas9 system has been extensively applied for crop improvement. However, our understanding of Cas9 specificity is very limited in Cas9-edited plants. To identify on- and off-target mutation in an edited crop, we described whole genome sequencing (WGS) of 14 Cas9-edited cotton plants targeted to three genes, and three negative (Ne) control and three wild-type (WT) plants. In total, 4188-6404 unique single-nucleotide polymorphisms (SNPs) and 312-745 insertions/deletions (indels) were detected in 14 Cas9-edited plants compared to WT, negative and cotton reference genome sequences. Since the majority of these variations lack a protospacer-adjacent motif (PAM), we demonstrated that the most variations following Cas9-edited are due either to somaclonal variation or/and pre-existing/inherent variation from maternal plants, but not off-target effects. Of a total of 4413 potential off-target sites (allowing <= 5 mismatches within the 20-bp sgRNA and 3-bp PAM sequences), the WGS data revealed that only four are bona fide off-target indel mutations, validated by Sanger sequencing. Moreover, inherent genetic variation of WT can generate novel off-target sites and destroy PAMs, which suggested great care should be taken to design sgRNA for the minimizing of off-target effect. These findings suggested that CRISPR/Cas9 system is highly specific for cotton plants.