Login / Signup

Derivation and validation of an epigenetic frailty risk score in population-based cohorts of older adults.

Xiangwei LiThomas DelerueBen SchöttkerBernd HolleczekEva GrillAnnette PetersMelanie WaldenbergerBarbara ThorandHermann Brenner
Published in: Nature communications (2022)
DNA methylation (DNAm) patterns in peripheral blood have been shown to be associated with aging related health outcomes. We perform an epigenome-wide screening to identify CpGs related to frailty, defined by a frailty index (FI), in a large population-based cohort of older adults from Germany, the ESTHER study. Sixty-five CpGs are identified as frailty related methylation loci. Using LASSO regression, 20 CpGs are selected to derive a DNAm based algorithm for predicting frailty, the epigenetic frailty risk score (eFRS). The eFRS exhibits strong associations with frailty at baseline and after up to five-years of follow-up independently of established frailty risk factors. These associations are confirmed in another independent population-based cohort study, the KORA-Age study, conducted in older adults. In conclusion, we identify 65 CpGs as frailty-related loci, of which 20 CpGs are used to calculate the eFRS with predictive performance for frailty over long-term follow-up.
Keyphrases
  • dna methylation
  • community dwelling
  • genome wide
  • risk factors
  • gene expression
  • physical activity
  • peripheral blood
  • machine learning
  • deep learning
  • copy number
  • genome wide association study
  • drug induced