# TGAge: A Phenotypic Aging Clock Based on Tabular Foundation Models and the Gompertz Law for Assisted Living Facility Residents

Preprint · Chengzhuo Li, Rongjia Zhou, Shanshan Wang, Runze Yan, Carl Yang, Xiaobo Sun, Huiwen Xu, Jiaying Lu

- Archive id: arxiv:2610.32994
- Published: 2026-09-30
- arXiv: https://arxiv.org/abs/2610.32994
- Citations: 0
- Source URL: https://arxiv.org/abs/2610.32994
- PDF: https://arxiv.org/pdf/2610.32994
- Provenance: arxiv

## Abstract

Biological age captures overall health status beyond chronological age, yet existing biological aging clocks are difficult to apply to assisted living facility (ALF) residents due to scarce epigenetic data. We propose TGAge, a second-generation phenotypic aging clock that estimates risk-equivalent biological age from routinely collected clinical measurements. TGAge employs a tabular foundation model to predict all-cause mortality risks over discrete time-to-death intervals and converts the resulting mortality-risk ratio into biological age acceleration via the Gompertz law. We develop and evaluate TGAge on a large cohort of ALF residents curated from a proprietary, multi-state long-term care database. TGAge outperforms clinical indices and established biological aging clocks in survival prediction. Moreover, TGAge-derived age acceleration is significantly associated with multiple aging-related comorbidities. These findings support TGAge as a practical, interpretable, and highly predictive biological aging clock for vulnerable ALF populations using only routinely collected clinical measurements.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [Biology of Aging & Longevity](https://space-kitty.com/topics/aging-longevity)

Cited in dossiers:
- [Biology of Aging & Longevity: State of the Field (October 2026)](https://space-kitty.com/dossiers/aging-longevity)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.