Navigating the CRISPR-Cas9 Frontier: AI-Enabled off-target prediction and sgRNA Design for Unprecedented Precision

Abstract

CRISPR-Cas9, composed of Clustered Regularly Interspaced Short Palindromic Repeats and CRISPR-associated protein 9, is a pivotal tool for precise genetic manipulation with diverse biomedical applications. In this system, the scientists use single guide RNA (sgRNA) synthesized from tracrRNA and crRNA to lead the Cas9 protein to target specific gene locations, thereby achieving gene editing. Nonetheless, drawbacks like single guide RNA's limited efficiency led to frequent base mismatch and off-target effects, which hampers CRISPR-Cas9's potential. Under these circumstances, entering machine learning, adept at adapting to variations and handling intricate datasets, is a viable avenue for optimizing CRISPR-Cas9's guide RNA by rectifying these limitations. Nevertheless, machine learning is not exempt from limitations. Within this framework, this paper presents a succinct overview of the challenges linked to sgRNA's efficiency issues. It then outlines existing mechanisms in machine learning and assesses the efficacy of machine learning in enhancing sgRNA design to improve CRISPR/Cas9 sgRNA specificity. Additionally, this paper scrutinizes notable restrictions and suggestions of machine learning in the quest for superior sgRNAs to guide future research.