Daneshpajouh, A.; Fowler, M.; Wiese, K. C.

Genome editing has the potential to treat genetic disorders at the source. This can be achieved by modifying the defective DNA through the intentional insertion, deletion, or substitution of genomic content. Among all genome editing technologies, CRISPR/Cas (Clustered Regularly Interspaced Short Palindromic Repeats and CRISPR-associated protein) is considered the gold standard. CRISPR/Cas uses a single guide RNA (sgRNA) to direct the Cas nuclease to a target DNA region. The efficacy of the complex depends on the ability of the sgRNA to bind to a complementary DNA sequence, which varies based on the sequence. A major challenge is identifying sgRNAs with high efficacy. We use a large dataset to build and compare several machine learning architectures for predicting on-target CRISPR/Cas activity, and explore the effect of including GC content as a biological feature. Our novel hybrid model, ChromeCRISPR, combines Convolutional Neural Networks (CNN) with Recurrent Neural Networks (RNN), and outperforms state-of-the-art models, including DeepHF and AttCRISPR, setting new benchmarks for predictive accuracy in CRISPR/Cas9 efficacy predictions.