Our bodies follow a set of genetic instructions. When those instructions get messed up, cells can grow out of control and turn into cancer.
Why Chromosome Problems Matter
One of the first signs of trouble is a change in a cell’s chromosomes. This can mean the wrong number of chromosomes or pieces that are broken or rearranged. Such defects push healthy cells toward becoming cancerous.
New AI Tool From EMBL
Scientists at the Korbel Group in Heidelberg created an AI‑based system that spots these chromosome errors fast. The tool helps researchers see how and when the mistakes happen, giving clues about the start of cancer.
Old Idea, New Technology
More than a hundred years ago, Theodor Boveri suggested that abnormal chromosomes might cause cancer. Yet, finding those cells was hard because they are rare and often die quickly. Researchers had to look under a microscope one cell at a time.
Introducing MAGIC
Marco Cosenza and his team built a platform that combines microscopy, single‑cell sequencing, and AI. They named it MAGIC – Machine learning‑Assisted Genomics and Imaging Convergence.
How MAGIC Works Like Laser Tag
MAGIC first takes thousands of pictures of cells with an automated microscope. An AI model, trained on examples of cells that contain a tiny structure called a micronucleus, scans the images.
A micronucleus is a small pocket that holds bits of DNA separate from the main genome. Cells with micronuclei are more likely to develop further chromosome errors.
When the AI spots a micronucleus, it tells the microscope to aim a laser at that exact cell. The laser triggers a special dye that changes color, permanently marking the cell.
Collecting and Studying Tagged Cells
After marking, scientists can pull out the dyed cells using flow cytometry. Those cells are then examined in detail, including a deep look at their DNA.
Because MAGIC does the work automatically, it can analyze almost 100,000 cells in a single day – far more than a person could ever check by hand.
What the Researchers Learned
Using MAGIC on normal human cells, the team found that about 10% of cell divisions create a spontaneous chromosome mistake. When the tumor‑suppressor gene p53 is mutated, that rate almost doubles.
The scientists also studied how the location of DNA breaks affects error formation.
Future Possibilities
MAGIC is flexible. Although it was trained to find micronuclei, the same AI can be taught to look for many other visual features inside cells.
"If you can see a difference under the microscope, the AI can learn to spot it," said Korbel. "That means this system could help many areas of biology, not just cancer research."