Key Points
- Scientists created a new algorithm, MultistageOT, to model how cells develop over time.
- Studying this process is difficult because current methods destroy the cells under observation.
- The new algorithm can reconstruct the entire developmental process from a single “snapshot.”
- It can predict how a cell will mature and can spot cells that are developing abnormally.
Scientists have developed a new computer algorithm that can predict how our cells develop and specialize, a breakthrough that could help us better understand how diseases like cancer begin.
The process by which a stem cell differentiates into a specific cell type, such as a neuron or an immune cell, is called cell differentiation. It’s a fundamental process of life, but studying it is incredibly difficult.
The main problem is that when scientists analyze a single cell to see what it’s doing, the process destroys the cell. This means they only get one “snapshot” in time and can’t see what the cell would have become.
To address this, researchers at Sweden’s Karolinska Institutet and KTH have developed a new algorithm, MultistageOT, that uses advanced mathematical principles to reconstruct the entire developmental trajectory from a single snapshot of a group of cells, thereby filling in missing steps.
“Our method makes it possible to model the entire developmental process, even if the cells are observed at a single time point,” said Magnus Tronstad, a doctoral student and author of the study.
The algorithm can predict how cells will mature and what their final function will be. In a test, the researchers used it to model the complex development of blood cells, and the results were a success. The tool was even able to identify cells that deviated from the normal path, which is crucial for detecting when processes go awry and lead to disease.
