Unveiling neural network potential in forecasting CRISPR effects and off-target prophecies for gene editing

Abstract

The revolutionary Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) - associated protein 9 (Cas9) systems has emerged as a groundbreaking gene-editing tool, widely embraced within biomedical research. Nonetheless, the utilization of guide RiboNucleic Acids (gRNAs) in the CRISPR-Cas9 system can inadvertently trigger undesired off-target effects, consequently impinging on the practical implementation of this technique. Existing in silico prediction methods that focus on off-target effects have exhibited constrained predictive accuracy, necessitating further enhancement. To tackle this challenge, we present a Base Editing and Prime Editing approach in this study. This approach aim to enhance the precision and specificity of DeoxyRiboNucleic Acid (DNA) modifications compared to traditional CRISPR-Cas9 methods. Both techniques provide unique approaches to achieve targeted changes in the DNA sequence without inducing double-stranded breaks, which can lead to off-target effects. Base editing is highly specific and allows for the correction of point mutations associated with diseases while Prime Editing allows for a wider range of modifications compared to base editing, including the ability to insert or delete specific sequences. Their ability to achieve specific genetic changes while minimizing off-target effects makes them valuable additions to the gene editing toolkit. The findings of this research contribute to the advancement of precision gene editing, offering an enhanced predictive framework to mitigate off-target effects in the realm of CRISPR-Cas9 technology.