CRISPR-Cas9 gRNA efficiency prediction - An overview of predictive tools and the role of Deep Learning

Summary

The CRISPR-Cas9 system has revolutionized the field of genome editing and promises the ability to study genetic interactions at their origin and the opportunity to cure severe inherited diseases. Compared with previous gene-editing tools, such as zinc-finger nucleases (ZFNs) and transcription activator-like effector nucleases (TALENs), which bind to specific DNA sequence by protein-DNA recognition, the CRISPR-Cas9 system identifies target sites by the complementarity between the guide RNA (gRNA) and the DNA sequence, which is less ex- pensive and time-consuming, as well as more precise and scalable. However, low cleavage efficiency and off-target effects hamper the development and application of CRISPR-Cas systems. To predict cleavage efficiency and specificity, numerous computational approaches have been developed. Nonetheless, currently available tools cannot robustly predict experimental success as prediction accuracy depends on the assumptions of the underlying model and how closely the experimental setup matches the training data. Moreover, new deep learning tools have been explored lately for gRNA efficiency prediction but have not been systematically evaluated. In this study, we present the approaches that pertain to the on-target activity problem, focusing mainly on the features and computational methods they utilize. Further- more, we evaluate these tools on independent datasets and give some suggestions for their usage. Finally, based on this analysis, we introduce a new gRNA design tool, named CRISPRedict, that provides accurate and interpretable predictions which can guide genome editing experiments and make plausible hypotheses for further investigation. CRISPRedict is available for use at http://www.crispredict.org/.