A domestic research team developed a new artificial intelligence (AI) technology that mathematically models cells and drug responses by breaking them down and reassembling them like Lego blocks, enabling it to predict not only new reactions between cells and drugs that have not been experimentally tested but also arbitrary gene regulation effects.
A research team led by Professor Cho Kwang-hyun of the Department of Bio and Brain Engineering at Korea Advanced Institute of Science and Technology (KAIST) said on the 16th that it developed a new AI technology that uses generative AI to identify drugs and gene targets that can guide cells to a desired state. The study was published on the 15th in the international journal Cell Systems, published by Cell.
Adjusting the state of cells in a desired direction is a core task in the life sciences, including new drug development, cancer treatment, and regenerative medicine, but finding suitable drugs or gene targets is not easy.
The team devised a method that, in an invisible, map-like space where image-generating AI mathematically organizes the features of objects or cells, separates the state of cells and the effects of drugs, recombines them, and predicts the responses of untested cell–drug combinations. Extending this principle, they were also able to predict what changes would appear when a specific gene was regulated.
They then validated the technology using real-world data. As a result, the AI identified molecular targets capable of reverting colorectal cancer cells to a state closer to normal cells, which was verified in cell experiments.
This demonstrates that the results are not limited to cancer treatment but constitute a general-purpose platform that can predict a wide range of untrained cell state transitions and drug responses. In other words, it is significant not merely at the level of "this drug is effective," but in that it could also reveal the mechanisms by which the drug acts inside cells. It is expected to be widely used in various medical fields in the future, not only for new drug development or cancer treatment but also for research that revives damaged cells to function like healthy ones.
Professor Cho said, "Inspired by image-generating AI technology, we applied the concept of a 'direction vector,' an idea that cells can also be changed in a desired direction," adding, "This technology is significant as a general-purpose AI approach that can quantitatively analyze the effects of specific drugs or genes on cells and even predict unknown responses."
References
Cell Systems (2025), DOI: https://doi.org/10.1016/j.cels.2025.101405