Abstract
CRISPR is considered a powerful tool for targeted genome editing. However, off-target effects remain a significant challenge in the CRISPR field, hindering its broader clinical application. To enhance the development of gene-editing therapies, it is essential to predict the efficiency of CRISPR-based genome editing experiments, before trying them on clinical cases. Machine learning (ML) and deep learning (DL) tools are projected to become the leading methods for predicting CRISPR on-target and off-target activity. Current prediction accuracy is limited by the amount of available training data. As more sequence features are identified and incorporated in DL tools, predictions of them are expected to better align with experimental results. Hence, the increasing focus on ML/DL approaches to predict off-target sites necessitates large and easily searchable databases. In this review, we will take a closer look at available CRISPR databases.