Identification of CRISPR/Cas9 off-target sites through the computational analysis of GUIDE-seq data

The CRISPR-Cas9 system is a genome editing tool derived from prokaryotes that became widely used in the past few years for the generation of site-specific genomic modifications in cell lines or even entire organisms. Although this system is programmed to recognize and to cut a specific DNA target sequence, it can also cleave off-target regions in the genome that present a similar DNA sequence. The cleavage of off-target regions can occur at a higher or lower frequency rate compared to on-target regions and could generate unwanted effects that need to be carefully considered when designing gene therapy strategies based on the CRISPR-Cas9 technology. Therefore, before its application in clinical trials, it is essential to investigate and determine the safety of the CRISPR-Cas9 system in terms of number and frequency of off-target events. The genome-wide unbiased identification of double-strand breaks (DSBs) enabled by sequencing (GUIDE-seq) is a new in vitro approach to detect the genome editing activity of the CRISPR-Cas9 system. This technique allows the genome-wide identification of off-target sites through the integration of double stranded oligodeoxynucleotide (dsODNs) into DSBs within the genome by non-homologous end joining (NHEJ). The selective amplification and sequencing of only those genomic fragments with integrated dsODNs allow to identify CRISPR-Cas9 cleavage sites genome-wide. In this study, we performed a bioinformatics analysis of GUIDE-seq data using two different tools, the †œguideseq†� Python package and the †œGUIDEseq†� R/Bioconductor package, for the genome-wide detection and annotation of on- and off-target cleavage sites. We found that the †œguideseq†� Python package is able to detect more off-target sites than the †œGUIDEseq†� R/Bioconductor package, although the majority of off-target sites are shared between two packages. Moreover, we analyzed publicly available histone modification profiles to evaluate the chromatin context in which the cleavage of off-target regions occurred. In fact, different cell types have different chromatin backgrounds that reflect cell type specific transcriptional programs and biological functions. The definition of chromatin states in different cell types allows to understand if different chromatin contexts could interfere with the ability of the CRISPR/Cas9 system to induce DSBs and if the off-target cuts fall in functionally relevant genomic regions such as enhancers or actively transcribed genes. We observed that the cleavage of most off-target sites occurs in quiescent chromatin regions devoid of any histone modification that characterize active regulatory elements. These results reveal that the use of epigenomic data is an effective approach to better characterize the impact of CRISPR/Cas9 off-targets.