Finding drugs that clear them normally means testing hundreds of thousands of compounds by hand. Researchers at the University of Edinburgh trained a machine learning model on published senolytic data instead, ranking candidates before a single new compound was tested in a dish.
The model screened a large virtual library and flagged a short list it predicted would clear senescent cells. In lab tests, three of them worked: ginkgetin, periplocin and oleandrin killed senescent human cells while sparing healthy ones, at a fraction of the usual screening cost.
Two of the three, periplocin and oleandrin, belong to a known family of heart drugs already suspected of senolytic activity, which cross-checked the model rather than simply repeating known results. The team estimated the approach cut screening costs more than 200-fold. None of the three compounds has been tested in people for aging, and the work stopped at cells in a dish.
Read the original study
Smer-Barreto et al., Nature Communications, 2023 · doi.org/10.1038/s41467-023-39120-1
A computer did the chemists' screening work.
Source
- Smer-Barreto, V., Quintanilla, A., Elliott, R. J. R., Dawson, J. C., Sun, J., Campa, V. M., Lorente-Macías, Á., Unciti-Broceta, A., Carragher, N. O., Acosta, J. C., & Oyarzún, D. A. (2023). Discovery of senolytics using machine learning. Nature Communications, 14(1), 3445. https://doi.org/10.1038/s41467-023-39120-1