Single-cell senescence identification reveals senescence heterogeneity, trajectory, and modulators

Area of Science:

• Gerontology
• Computational Biology
• Cell Biology

Background:

• Cellular senescence is implicated in aging and disease, but its complexity hinders research.
• Heterogeneity of senescent cells presents challenges for accurate identification and targeted therapies.

Purpose of the Study:

• To develop and validate a machine learning program, SenCID, for precise identification of senescent cells.
• To characterize distinct senescence identities (SIDs) and their functional implications.
• To enable the analysis of senescence trajectories in various physiological and pathological conditions.

Main Methods:

• Machine learning model (SenCID) trained on 602 transcriptome datasets from 52 senescence studies.
• Analysis of bulk and single-cell transcriptomic data.
• Application to Perturb-seq data to identify senescence modulators.

Main Results:

• SenCID accurately identifies senescent cells across diverse cell types and datasets.
• Six major senescence identities (SIDs) were identified, each with unique characteristics.
• SenCID revealed differential responses of SIDs to senolytic treatments.
• Senescence trajectories were reconstructed in aging, chronic diseases, and COVID-19.
• Senescence modulators were identified using single-cell Perturb-seq data.

Conclusions:

• SenCID is a powerful tool for precise single-cell analysis of cellular senescence.
• Understanding senescence heterogeneity through SIDs can guide targeted therapeutic interventions.
• SenCID facilitates research into aging-related pathologies and disease mechanisms.