The accumulation of senescent cells is increasingly recognized as a hallmark of biological aging, linked to various chronic medical conditions. Senescent cells are known to exhibit a unique secretory profile, termed the senescence-associated secretory phenotype (SASP), which can influence surrounding and distant tissues. This study focuses on the development and validation of a composite SASP Score that integrates large-scale population proteomics data with a semi-supervised deep learning framework.
The analytical workflow involved several key steps: first, the selection of biologically curated SASP proteins; second, the development of a Guided autoencoder with Transformer (GAET) model using data from the UK Biobank Pharma Proteomics Project (UKB-PPP); third, internal evaluation and association analyses within the UK Biobank; and finally, external validation and longitudinal assessment in an independent randomized clinical trial cohort.
The results indicated that the deep learning-based SASP Score serves as a strong, independent predictor of mortality risk and the onset of serious chronic medical conditions such as dementia, chronic obstructive pulmonary disease (COPD), myocardial infarction, and stroke. Notably, in an independent cohort, multimodal exercise was shown to significantly alter the SASP Score trajectory over an 18-month period.
These findings suggest that the SASP Score could be a valuable biomarker for assessing systemic cellular senescence burden. Moreover, the statistical approach employed in this study enhances interpretability and cross-platform utility, making it a promising tool for aging research and the evaluation of interventions guided by geroscience.
What this study does not show
This study does not establish a direct causal relationship between changes in the SASP Score and specific health outcomes, nor does it provide evidence of the clinical utility of the SASP Score in guiding treatment decisions.
Evidence scorecard
- Analytic validity: Established
- Clinical validity: Partial
- Clinical utility: Not established
The evidence is based on primary research involving human participants.