Highly Parallel Profiling of the Activities and Specificities of Cas12a Variants in Human Cells

Several Cas12a variants have been developed to broaden its targeting range, improve the gene editing specificity or the efficiency. However, selecting the appropriate Cas12a among the many orthologs for a given target sequence remains difficult. Here, we perform high-throughput analyses to evaluate the activity and compatibility with specific PAMs of 24 Cas12a variants and develop deep learning models for these Cas12a variants to predict gene editing activities at target sequences of interest. Furthermore, we reveal and enhance the truncation in the integrated tag sequence that may hinder off-targeting detection for Cas12a by GUIDE-seq. This enhanced system, which we term enGUIDE-seq, is used to evaluate and compare the off-targeting and translocations of these Cas12a variants.