Researchers at EMBL Heidelberg developed MAGIC to study chromosomal abnormalities linked to the early stages of cancer development, leveraging machine learning-assisted genomics and imaging convergence.

Additional findings indicate that roughly 10% of cell divisions yield chromosomal abnormalities, increasing notably with p53 mutation, while other factors such as double-stranded DNA breaks influence instability.

In particular, the technique supports high-throughput, single-cell analysis, speeding up the detection of chromosomal abnormalities by processing large cell cohorts quickly.

The Nature paper reflects a collaborative effort across EMBL facilities and external partners, including EMBL-EBI and DKFZ, as part of the Molecular Medicine Partnership Unit with the University of Heidelberg.

A new AI-powered system called MAGIC automates detection of micronuclei in cells using automated microscopy, a trained machine learning model, and a photoconvertible dye to tag targeted cells for deeper genomic analysis.

the study finds that just over 10% of cell divisions produce spontaneous chromosomal abnormalities, and the rate nearly doubles when the tumor suppressor gene p53 is mutated, with double-stranded DNA breaks also examined as a trigger.

MAGIC is adaptable and trainable to identify other visually distinguishable cellular features, suggesting broad potential applications in biology.

The results reinforce that chromosomal abnormalities are a major driver of aggressive cancers and may illuminate the early development and progression of cancer.

The MAGIC approach enables high-throughput analysis, allowing researchers to examine nearly 100,000 cells in a single day, dramatically accelerating studies of early cancer origins.

The study's publication in Nature on October 29, 2025, highlights broad collaboration with institutions like DKFZ, EMBL-EBI, and MMPU.

The work builds on historic links between chromosomal abnormalities and cancer and aims to generalize the approach to detect other visually discernible cellular features in various biological contexts.

Micronuclei-containing cells are tagged and isolated (for example by flow cytometry) to enable deeper genomic analyses and help reveal mechanisms underlying the earliest steps of cancer formation.