Home · Science · Sep 20 archive
UC San Diego Researchers Build Virtual Cells to Speed Drug Discovery
Developing
In Short: UC San Diego researchers have developed virtual cells using artificial intelligence (AI) models and 'digital twins' to predict how mitochondria in human cells will respond to drug treatments, potentially speeding up drug discovery for diseases like cancer, diabetes, and Alzheimer’s.
The team’s findings, published in Cell, show that virtual mitochondrial networks respond to drugs in a manner closely matching real cells, reducing the need for time-consuming lab experiments.
One approach involved training a deep-learning AI model on 40,000 4D movies of drug-treated cells to predict cellular health based on mitochondrial shape alone.
The other approach created a physics-based digital twin of a living cell, defining rules about how its organelles behave and implementing them in a model.
To test the model, the team used an advanced imaging technique to create high-fidelity 4D movies of mitochondria within human cells responding to drug treatment.
Without knowing which drug was used on each cell, the model produced an organized map that grouped cells with similar responses together, demonstrating its predictive power.
The researchers treated cancer cells with 25 different compounds known to perturb mitochondria, producing 40,000 single-cell 4D movies.
According to Schöneberg, the team has built a physics-based virtual cell that can be compared side-by-side with actual 3D microscopy movies, a first.
These studies could significantly reduce the time and resources required for drug discovery, aligning with the FDA’s efforts to speed up the drug approval process.
Background
The Food and Drug Administration (FDA) has opened applications for a pilot program aimed at reducing the time it takes for drugmakers to begin human testing, as part of efforts to compete with global biotech industries.
What's still developing
- Unlike most AI models that require humans to manually label images, MitoSpace found patterns on its own, learning what makes the mitochondria of one cell distinct from that of another.
- “We have built a physics‑based virtual cell and can compare it side‑by‑side to the actual 3D microscopy movie, something that has never been possible before,” said Schöneberg.
- To test the model, the team asked it to predict how mitochondria would respond when microtubules were partially broken down by a drug called nocodazole.
- Now, researchers at University of California San Diego have employed two different approaches to quantify these morphological changes by creating “virtual cells” — digital models that mimic the dynamic biological processes of real cells.