Massively parallel CRISPR off-target detection enables rapid off-target prediction model building

BACKGROUND

CRISPR (clustered regularly interspaced short palindromic repeats) genome editing holds tremendous potential in clinical translation. However, the off-target effect has always been a major concern.

METHODS

Here, we have developed a novel sensitive and specific off-target detection method, AID-seq (adaptor-mediated off-target identification by sequencing), that can comprehensively and faithfully detect the low-frequency off targets generated by different CRISPR nucleases (including Cas9 and Cas12a).

FINDINGS

Based on AID-seq, we developed a pooled strategy to simultaneously identify the on/off targets of multiple gRNAs, as well as using mixed human and human papillomavirus (HPV) genomes to screen the most efficient and safe targets from 416 HPV gRNA candidates for antiviral therapy. Moreover, we used the pooled strategy with 2,069 single-guide RNAs (sgRNAs) at a pool size of about 500 to profile the properties of our newly discovered CRISPR, FrCas9. Importantly, we successfully built an off-target detection model using these off-target data via the CRISPR-Net deep learning method (area under the receiver operating characteristic curve [AUROC] = 0.97, area under the precision recall curve [AUPRC] = 0.29).

CONCLUSIONS

To our knowledge, AID-seq is the most sensitive and specific in vitro off-target detection method to date. And the pooled AID-seq strategy can be used as a rapid and high-throughput platform to select the best sgRNAs and characterize the properties of new CRISPRs.

FUNDING

This work was supported by The National Natural Science Foundation of China (grant nos. 32171465 and 82102392), the General Program of Natural Science Foundation of Guangdong Province of China (grant no. 2021A1515012438), Guangdong Basic and Applied Basic Research Foundation (grant no. 2020A1515110170), and the National Ten Thousand Plan-Young Top Talents of China (grant no. 80000-41180002).

A systematic evaluation of data processing and problem formulation of CRISPR off-target site prediction

CRISPR/Cas9 system is widely used in a broad range of gene-editing applications. While this editing technique is quite accurate in the target region, there may be many unplanned off-target sites (OTSs). Consequently, a plethora of computational methods have been developed to predict off-target cleavage sites given a guide RNA and a reference genome. However, these methods are based on small-scale datasets (only tens to hundreds of OTSs) produced by experimental techniques to detect OTSs with a low signal-to-noise ratio. Recently, CHANGE-seq, a new in vitro experimental technique to detect OTSs, was used to produce a dataset of unprecedented scale and quality (>200 000 OTS over 110 guide RNAs). In addition, the same study included in cellula GUIDE-seq experiments for 58 of the guide RNAs. Here, we fill the gap in previous computational methods by utilizing these data to systematically evaluate data processing and formulation of the CRISPR OTSs prediction problem. Our evaluations show that data transformation as a pre-processing phase is critical prior to model training. Moreover, we demonstrate the improvement gained by adding potential inactive OTSs to the training datasets. Furthermore, our results point to the importance of adding the number of mismatches between guide RNAs and their OTSs as a feature. Finally, we present predictive off-target in cellula models based on both in vitro and in cellula data and compare them to state-of-the-art methods in predicting true OTSs. Our conclusions will be instrumental in any future development of an off-target predictor based on high-throughput datasets.