A University of Zurich-led team developed 'Pythia', an AI tool for designing microhomology-based DNA repair templates that achieve highly precise and predictable CRISPR/Cas9 genome integrations. Published in Nature Biotechnology, this method enhances gene editing accuracy by forecasting cellular repair outcomes based on sequence context, reducing unintended alterations often seen with end joining repair pathways. The technology demonstrated efficacy across multiple loci in human cells and animal models, promising safer gene therapies and disease models.