Related Experiment Video
Updated: Jun 2, 2026
05:53
Measuring Frailty in HIV-infected Individuals. Identification of Frail Patients is the First Step to Amelioration and Reversal of Frailty
Published on: July 24, 2013
A Preliminary Predictive Panel for Pre-Frailty Based on Serum Proteomic Biomarkers: A Two-Phase Cross-Sectional Study
Yu Ye1,2, Jinwei Liu1,2, Zhen Zhang1,2
1Medical School, Chinese PLA General Hospital, Beijing, People's Republic of China.
Clinical Interventions in Aging
|June 1, 2026
Summary
This study identified novel serum protein biomarkers for prefrailty, a condition linked to protein metabolism issues. Machine learning models using proteins like UBE2O and ACTN3 show promise for early prefrailty detection in older adults.
Area of Science:
- Gerontology
- Proteomics
- Biomarker Discovery
Background:
- Prefrailty is linked to protein metabolism dysfunction, but its serum proteomic signature is not well understood.
- Identifying prefrailty early is crucial for intervention in older adults.
Purpose of the Study:
- To investigate serum protein profiles across different frailty statuses in older adults.
- To identify potential serum biomarkers for early prefrailty detection.
Main Methods:
- Untargeted proteomic screening followed by parallel reaction monitoring (PRM) validation in a cohort of older adults.
- Integration of clinical parameters and differentially expressed proteins into machine learning models.
- Categorization of participants into robust, prefrail, and frail groups.
Main Results:
- 166 differentially expressed proteins were identified across frailty statuses; 15 prefrailty-associated proteins were validated by PRM.
- Enrichment analysis revealed involvement in cell signaling, protein metabolism, immune regulation, and muscle function.
- A Random Forest model incorporating gait speed, muscle mass, UBE2O, Timed Up and Go, ACTN3, and MMSE showed robust performance for prefrailty identification.
Conclusions:
- Candidate serum protein biomarkers associated with prefrailty were identified.
- Preliminary machine learning models incorporating UBE2O and ACTN3 demonstrated feasibility in discriminating prefrailty from robust status.
- Findings suggest underlying proteomic heterogeneity in community-dwelling older adults and potential for early prefrailty detection.
