Summarized by Masters of Longevity from Fight Aging!.
Reports a convolutional neural network that predicts hematopoietic stem cell chromatin age from 3D nuclear images and highlights interpretable chromatin features linked to aging.
Key Takeaways
- Researchers developed ChromAgeNet, a CNN that discriminates young versus aged murine HSC nuclei with AUROC 0.77 ± 0.03.
- Explainable AI identified chromatin entropy, peripheral heterochromatin, and chromatin condensates as the model's predictive features.
- The model detected epigenetic drug–induced shifts in aged HSC chromatin and is proposed as a phenotypic screening tool for rejuvenation.
