The accumulation of senescent cells is increasingly recognized as a hallmark of biological aging, associated with various chronic medical conditions. Senescent cells are characterized by a distinct secretory profile known as the senescence-associated secretory phenotype (SASP). This profile can propagate cellular senescence to neighboring and distant tissues, making it a significant focus for aging research.
In this study, researchers developed and validated a composite SASP Score by integrating large-scale population proteomics data with a semi-supervised deep learning framework. The analytical workflow included several key steps:
- Selection of biologically curated SASP proteins: This step involved identifying proteins known to be associated with the SASP.
- Development of a Guided autoencoder with Transformer (GAET) model: This model was developed using data from the UK Biobank Pharma Proteomics Project (UKB-PPP).
- Internal evaluation and association analyses: The model was evaluated within the UK Biobank to assess its predictive capabilities.
- External validation and longitudinal assessment: The SASP Score was validated in an independent randomized clinical trial cohort.
The findings indicated that the deep learning-based SASP Score was a strong, independent predictor of mortality risk and the onset of serious chronic medical conditions, including dementia, chronic obstructive pulmonary disease (COPD), myocardial infarction, and stroke. Furthermore, in an independent cohort, multimodal exercise significantly altered the trajectory of the SASP Score over an 18-month period.
These results support the potential of the deep learning-derived SASP Score as a biomarker for systemic cellular senescence burden. The statistical approach employed in this research offers enhanced interpretability and cross-platform utility, positioning the SASP Score as a valuable tool for aging research and the evaluation of geroscience-guided interventions.
What this study does not show
While the study identifies the SASP Score as a predictor of mortality and chronic conditions, it does not establish a direct causal relationship between SASP levels and health outcomes.
Evidence scorecard
- Analytic validity: Established
- Clinical validity: Partial
- Clinical utility: Partial
The evidence is based on primary research involving human subjects.