Researchers at the Duke University School of Medicine have developed a powerful new method called DeepScence to accurately identify senescent cells, which are cells that have stopped dividing and can contribute to aging, tissue dysfunction, and disease progression. Senescent cells play a critical role in processes like inflammation, cancer development, and age-related disorders, but detecting them has been a major challenge due to their heterogeneous molecular profiles and spatial distribution within tissues.

DeepScence uses advanced deep neural networks combined with a curated gene set called CoreScence, which incorporates information from multiple published senescence-associated gene sets. By analyzing single-cell RNA sequencing and spatial transcriptomics data, DeepScence can identify senescent cells both in vitro, from lab-grown samples, and in vivo, directly within living tissues. Its flexibility across multiple platforms and high accuracy allows researchers to explore not only which cells are senescent, but also where they are located in tissue and how they interact with surrounding cells.

The method substantially outperforms existing approaches, enabling scientists to uncover patterns of senescence that were previously hidden. This is particularly important for understanding how senescent cells influence aging, tissue repair, and the progression of chronic diseases. By leveraging RNA sequencing data, DeepScence provides a detailed view of the molecular and spatial characteristics of senescent cells, helping researchers investigate potential interventions to target harmful senescent cells or support beneficial ones in regenerative medicine and anti-aging therapies.

Overall, DeepScence represents a significant advancement in cellular biology and computational analysis, giving scientists a reliable tool to map senescent cells at the single-cell and tissue level and to better understand their roles in health and disease.

Availability – The DeepScence Python package is freely accessible on GitHub (https://github.com/anthony-qu/DeepScence)