Micronuclei-containing cells are tagged and isolated, by methods such as flow cytometry, enabling deeper genomic analysis to reveal the earliest steps of cancer formation.

Collaborations across EMBL facilities and external groups (DKFZ, EMBL-EBI, MMPU) supported the study, with the Nature paper published on a late-October date in 2025.

The study finds that just over 10% of cell divisions spontaneously produce chromosomal abnormalities, with the rate nearly doubling when p53, a tumor suppressor gene, is mutated, alongside evaluation of other triggers like double-stranded DNA breaks.

The technique enables high-throughput, single-cell analysis, allowing nearly 100,000 cells to be studied in under a day, significantly accelerating detection of chromosomal abnormalities.

Researchers at EMBL Heidelberg developed an AI-based tool called MAGIC to study chromosomal abnormalities linked to early cancer development.

The approach builds on historical links between chromosomal abnormalities and cancer and aims to generalize to detect other visually discriminable cellular features in various biological contexts.

MAGIC automates micronuclei detection through automated microscopy, a trained machine learning algorithm, and a photoconvertible dye to tag targeted cells for further analysis.