Genome-wide prediction and profiling of Off-target cleavage by CRISPR=Cas9

Abstract

Although CRISPR RNA-guided nucleases (RGENs) are widely used in genome-editing studies, their off-target cleavage activity on a genome-wide scale has not been yet distinguished. Here, we developed a method to detect unintended DNA double-stranded breaks (DSBs) introduced by CRISPR-Cas9, and by extension, predict genome-wide off-target sites of specific guide RNA sequences through a deep learning-based method. This prediction tool implemented experimental in vivo data and thus, should be more accurate than previously reported computational prediction methods. Our effort represents the most exhaustive genome-wide survey of Cas9 off-target effects in evaluating its safety before clinical applications.