Google DeepMind and Yale University have created a new artificial intelligence system that can accelerate discoveries in the field of cancer treatment. The C2S-Scale 27B model can recognize complex biological patterns in cells and predict how different drugs affect tumors.
AI helped to clarify the mechanism that makes “cold” tumors visible to the immune system. Such tumors usually avoid recognition by the immune system, which complicates therapy, including immunotherapy. This discovery could pave the way for new combination treatments that were previously thought impossible.
C2S-Scale 27B analyzed more than 4,000 drugs on patient tumor samples and laboratory cells. He isolated substances that can selectively enhance the immune response without affecting all cells at once.
Among the key discoveries is the CK2 kinase inhibitor cilmitasertib (CX-4945). The model predicted that it works only when there is a small amount of interferon in the cells. When used separately, neither interferon nor the drug had a significant effect, but their combination increased the activity of the immune system by 50%, “heating up” tumors and making them visible to the body’s defense cells.
The system’s prediction was confirmed by laboratory tests on human neuroendocrine cells, with which the model had not previously worked. This shows that AI can not only process data, but also draw conclusions based on cellular context.
According to the researchers, large-scale AI models can become “virtual laboratories” that run thousands of simulations and discover new relationships between drugs, cells and the immune system.
“This discovery could open up a new path for developing cancer treatments”, said Google CEO Sundar Pichai.
The success of C2S-Scale 27B proves that AI can significantly accelerate scientific research and help create effective treatments faster than traditional approaches.
