Single-cell and spatial detection of senescent cells using DeepScence.

Area of Science:

• Cell Biology
• Bioinformatics
• Computational Biology

Background:

• Cellular senescence plays a critical role in aging and disease.
• Accurate identification of senescent cells is crucial for understanding their functions.
• Existing methods for senescent cell identification have limitations.

Purpose of the Study:

• To develop a robust and accurate computational method for identifying senescent cells.
• To leverage deep learning for analyzing single-cell and spatial transcriptomics data.
• To improve the characterization of senescent cells in various biological settings.

Main Methods:

• Development of DeepScence, a deep neural network-based method.
• Creation of CoreScence, a curated senescence-associated gene set.
• Application of DeepScence to single-cell and spatial transcriptomics datasets (in vitro and in vivo).

Main Results:

• DeepScence accurately identifies senescent cells in diverse transcriptomic datasets.
• The method demonstrates superior performance compared to existing approaches.
• Successful application across different experimental platforms for spatial transcriptomics.

Conclusions:

• DeepScence provides a powerful tool for senescent cell identification.
• This method facilitates deeper investigation into the spatial and molecular characteristics of senescent cells.
• DeepScence has broad applicability in aging and disease research.