Precise, predictable genome integrations by deep learning-assisted design of microhomology-based templates

This study aims to improve the precision of genome editing using the CRISPR-Cas9 system. Current gene editing approaches often result in unintended mutations due to unpredictable DNA repair. To address this, we developed a new method to guide the repair process by designing specific DNA sequences that encourage accurate integration of new genetic material. We tested this method in human cells, frog embryos (Xenopus tropicalis), and adult mouse brain tissue. The approach worked in both dividing and non-dividing cells, and also allowed for small, scarless edits at specific genomic locations. This research advances the field of genome editing by improving the reliability and accuracy of genetic modifications, with potential applications in basic research, gene therapy, and biotechnology.