An international team of researchers led by the University of Zurich (UZH) has developed a groundbreaking method for precise DNA modification. “Pythia” combines state-of-the-art genetic engineering with artificial intelligence (AI) and enables high-precision genome editing. It creates the basis for more accurate models of human diseases and future gene therapies, as the journal “Nature” reports.

Die CRISPR/Cas-Technologie, oft als „Genschere“ bezeichnet, ermöglicht gezielte Veränderungen der Erbinformation durch Punktmutationen oder die Integration ganzer Gene. Allerdings können doppelsträngige DNA-Brüche, die durch CRISPR ausgelöst werden, bei der Reparatur durch die Zelle unbeabsichtigte Fehler im Genom verursachen. Dies birgt Risiken, wie die Schädigung benachbarter Gene, was die Sicherheit und Wirksamkeit von Gentherapien einschränken kann.

The new Pythia method overcomes these problems by using AI that predicts DNA repair processes. The researchers developed tiny DNA repair templates that act like a molecular glue and guide the cell to make precise genetic changes. These templates were designed using machine learning that simulates millions of possible gene editing results. The AI analyzed consistent patterns in the DNA repair process to determine the most efficient pathway for targeted genetic modifications. The method was successfully tested in human cell cultures, in the tropical frog Xenopus and in brain cells of living mice, where it enabled high-precision gene modifications.

One advantage of Pythia is the ability not only to change individual bases of the genetic code, but also to insert larger genes – or to fluorescent label proteins to observe their function in healthy or diseased tissue. It is particularly noteworthy that the method also works in non-dividing cells, such as in the brain, which predestines it for neurological applications.

Original Paper:

Thomas Naert et al. Precise, predictable genome integrations by deep learning–assisted design of microhomology-based templates. Nature Biotechnology. 12 August 2025. DOI: 10.1038/s41587-025-02771-0

Editor: X-Press Journalistenbüro GbR

Gender Notice. The personal designations used in this text always refer equally to female, male and diverse persons. Double/triple naming and gendered designations are used for better readability. ected.