Optimization of CRISPR-Cas9 via the synergy of MD simulation and machine learning

CRISPR-Cas9, a promising gene-editing tool, sheds light on gene therapy. The normal DNA cleavage of CRISPR-Cas9 is programmed by a guide RNA (gRNA) template. However, recent studies showed that Cas9 cleavage occurs even without guidance from the gRNA in the presence of Mn2+ ions, implying the issue of off-target effect of Cas9. Here, we report a mechanism of this RNA-independent off-target cleavage (RI-cleavage) elucidated by molecular dynamic (MD) simulations. We further used machine learning algorithms developed by our lab to facilitate the design of novel Cas9 variants to reduce such RI-cleavage. In this study, we revealed the possible mechanism of RI-cleavage and further engineered Cas9 to reduce RI-cleavage via the power of machine learning. Our research serves as an excellent example showing the potential in the synergy of MD simulation and machine learning to optimize CRISPR-Cas9.