Discussion
CRISPR-Cpf1-based genome editing technology was invented as another promising tool following the success of CRISPR-Cas9 and have been applied for a broad spectrum of species including eukaryotic and prokaryotic organisms. Cpf1 has several advantages over Cas9 such as guided by a single and shorter RNA, requires a T-rich PAM and exhibits better specificity, making it a complementary and alternative gene editing tool.[37] However, predicting the activity of guide RNA is a challenge due to the lack of the knowledge.
In this study, we made a large-scale library containing 12544 designed guide RNAs .Take advantage of the library, we carry out screening experiments. and presented an efficient and comprehensive computational model for prediction of Cpf1 guide RNA activity trained by the results of which. Consequently, we systematically studied the guide RNA activity in bacteria for the first time, broadening the application of Cpf1 in genome editing. We observed various activities of guide RNAs and tested the results by independent validation experiments. Our trained model was developed to select optimized guide RNAs better-performing than previous approaches and analyzed the features contributed to guide RNA activities. We propose that deepCpf1 and other models trained in mammalian cells shows a higher false negative rate processing data from in prokaryotic organisms on account of misattributing influences of NHEJ repairs, chromatin structure, or lentivirus transduction to guide RNA incapacity, hence our model may be applied to other bacteria or unexplored host, potentially. Furthermore, our unbiased screening methods may be applied in other developing CRISPR system utilization scenario
However, there may be still several defects in our model that the mechanism of Cpf1-RNA complex interacting with DNA target is unclear and the generalization capacities are unproven. To further improve the accuracy and efficiency of guide RNA design, we expect that more datasets from well-designed experiments across species could be generated, and more comprehensive algorithms such as machine and deep learning (MDL) methods could be applied.[38-40]In addition, improvement of the specificity of guide RNA to reduce off-target effects is a clinical demand, thus our platform combining high-throughput screening and analysis of computational methods may be used to address this issue.
Currently, novel cpf1 orthologs and several other CRISPR nucleases were harnessed for genome editing in mammalian and bacteria cells, while relationship between guide RNA sequence and on-target activity was underexplored. [41]Moreover, several Cpf1 variants of AsCpf1 were described with increased activity and expanded protospacer-adjacent motif (PAM) preferences.[42,43]The method described in this paper could deal with such issues and make progress towards optimizing the design of guide RNAs.