Deep Learning Approaches for Off-targets Prediction in CRISPR-Cas9 Genome Editing to Improve Resistant in Plants

Genome editing enables the modification, insertion of mutations, and alteration of living organisms’ genomes. This breakthrough technology expands the possibilities for genetics, molecular biology, and biomedical research. CRISPR-Cas9 has revolutionized genome editing across various organisms, including plants. Deep learning techniques have been increasingly utilized for off-target prediction in CRISPR/Cas9 gene editing, aiming to enhance the accuracy and efficiency of identifying potential off-target sites. To analyze biological data and predict off-target sites with superior performance, I compared differed variants of deep learning algorithm. In this chapter, I used different variants of three deep learning models, namely feedforward neural network (FNN), convolutional neural network (CNN), and recurrent neural network (RNN) which significantly improved the prediction of off-target cleavage sites and genome vulnerability, achieving high accuracies of up to 99.5%. In three variants of FNN models, FNN5 outperforms from FNN3 and FNN7 with highest accuracy, low loss, good off-target and on-target prediction, and better F1 score. In two variants of CNN, CNN3 performs better than CNN5 in terms of overall evaluating parameters, and in two variants of RNN, RNN-GRU performs better than RNN-LSTM with high accuracy of 0.995, low loss of 0.0195, best off-target, and on-target prediction of CRISPR/Cas9. When compared the performance of all the discussed models, found that RNN-GRU outperforms all other models. The performance of the said models are evaluated based on several evaluation metrics such as confusion matrix, precision, recall, support, F1-Score, microaverage, macroaverage, and accuracy. Overall, the application of deep learning in off-target prediction for CRISPR-Cas9 gene editing showcases promising advancements in enhancing the precision and safety of genome editing techniques to develop resistance against various diseases in plants.