Researchers at the European Molecular Biology Laboratory (EMBL) have developed an AI-assisted system that could help scientists identify cancer causes early.
Developed by the intergovernmental agency supported by over 30 countries, Machine Learning-Assisted Genomics and Imaging Convergence (MAGIC) combines artificial intelligence, laser technology, and microscopy. Its creators at EMBL hope that one day, MAGIC will help bring about an understanding of the origins of cancer.
Cancer is a disease caused by out-of-control cell division, which can form tumors. This rapid cell division is caused by errors in a cell’s genetic code, which are essentially the instructions for how it functions. Chromosomal abnormalities, meaning “numerical and structural defects” in chromosomes, are known to be a cause of tumorous growth.
“Chromosomal abnormalities are a main driver for particularly aggressive cancers, and they're highly linked to patient death, metastasis, recurrence, chemotherapy resistance, and fast tumor onset,” according to Jan Korbel, senior scientist at EMBL.
The ability to detect these abnormalities is crucial to understanding how cancer works, but, until recently, extremely hard to do, since only a handful of healthy cells display chromosomal abnormalities at the same time, and are often killed off by the body’s natural selection process before they can be studied.
A MAGIC solution: Using AI to fight cancer
The notion that chromosomal abnormalities play a role in cancer growth isn’t recent, stretching back as far as the early 20th century. Despite this, MAGIC is the first system to effectively study the phenomenon.
MAGIC, which has been in development for several years and allows scientists to study the earliest genetic mistakes that lead to cancer, works almost like laser tag, albeit on a smaller scale.
First, a microscope takes a number of photos of dividing cells. Then, the AI portion of the system looks for micronuclei, small bubbles containing some of the cell’s DNA. Cells with micronuclei tend to produce chromosomal abnormalities, which is why the system specifically targets them.
Next, these selected cells are illuminated by a laser that uses a chemical dye to turn them a specific color, essentially marking them with a tag. Using methods such as flow cytometry, the tagged cells are separated and further studied.
In earlier times, cells had to be hand-checked, a process that took time and effort and offered little reward. MAGIC allows scientists to check over 100,000 cells in a single day, something that was unheard of previously.
Using MAGIC, the team at EMBL discovered that more than 10% of cell divisions result in chromosomal abnormalities. When p53, which regulates tumor growth, is suppressed, this number almost doubles.
The true value of MAGIC isn’t only its use for cancer research, but its versatility and ability to be used to detect all kinds of cellular features.
“As long as you have a feature that can be discriminated visually from a 'regular’ cell, you can – thanks to AI – train the system to detect it,” said Korbel, “Our system therefore has potential to advance future discoveries in numerous areas of biology.”