Virtual cells help researchers distinguish how drugs affect living cells

UC San Diego researchers built two computer models that track how mitochondria, the energy-producing structures inside cells, move and change shape after drug treatment. Both studies used three-dimensional microscopy movies to capture changes over time. An AI model trained on 40,000 movies grouped drugs by how they act with 75% accuracy, compared with 56% using flat images. A separate physics-based model reproduced the response of a real cancer cell's mitochondria to a drug without further adjustment. The studies, published in Cell, could help researchers test drug effects with fewer laboratory experiments. The team plans to combine the two approaches.