A computational solution for bolstering reliability of epigenetic clocks: Implications for clinical trials and longitudinal tracking

Abstract

Epigenetic clocks are widely used aging biomarkers calculated from DNA methylation data. Unfortunately, measurements for individual CpGs can be surprisingly unreliable due to technical noise, and this may limit the utility of epigenetic clocks. We report that noise produces deviations up to 3 to 9 years between technical replicates for six major epigenetic clocks. The elimination of low-reliability CpGs does not ameliorate this issue. Here, we present a novel computational multi-step solution to address this noise, involving performing principal component analysis on the CpG-level data followed by biological age prediction using principal components as input. This method extracts shared systematic variation in DNAm while minimizing random noise from individual CpGs. Our novel principal-component versions of six clocks show agreement between most technical replicates within 0 to 1.5 years, equivalent or improved prediction of outcomes, and more stable trajectories in longitudinal studies and cell culture. This method entails only one additional step compared to traditional clocks, does not require prior knowledge of CpG reliabilities, and can improve the reliability of any existing or future epigenetic biomarker. The high reliability of principal component-based epigenetic clocks will make them particularly useful for applications in personalized medicine and clinical trials evaluating novel aging interventions.

Effect of Long-Term Caloric Restriction on DNA Methylation Measures of Biological Aging in Healthy Adults: CALERIE™ Trial Analysis

“…Test-retest reliability ICC statistics for the aging measures were correlated with their treatment-effect magnitudes; the more-reliable measures had larger treatment effects. This result suggests that new methods proposed to improve reliability of DNAm measures of aging may reveal still-larger treatment effects in CALERIE and have potential to increase statistical power of other trials utilizing these measures as endpoints 50,51 . available measurements of this construct 53 .…”

“…We analyzed several proposed DNA methylation measures of aging, which are recognized as the best available measurements of this construct 53 . Nevertheless, these measures are acknowledged to be incomplete summaries of biological changes that occur with aging and to have technical limitations, including imperfect test-retest reliability 50,51,54,55 . Treatment-effect estimates may therefore represent a lower-bound of the true impact of CALERIE intervention on aging.…”

“…Our findings highlight two measures with potential utility in future trials, the GrimAge clock and the DunedinPACE Pace of Aging. These two measures have the highest test-retest reliability and show the strongest associations with healthspan endpoints in validation studies 25,28,51 . CALERIE trial evidence contributes further validation that these measures are appropriate for testing geroscience interventions in humans.…”

“…Our findings highlight two measures with potential utility in future trials, the GrimAge clock and the DunedinPACE Pace of Aging. These two measures have the highest test-retest reliability and show the strongest associations with healthspan endpoints in validation studies 25,28,51…”

A Multi-Omics Longitudinal Aging Dataset in Primary Human Fibroblasts with Mitochondrial Perturbations

“…DNAmAge computed from the combination of multiple probes or CpG sites showed a coefficient of variation, for selected clocks, of 3% for the Horvath1 (PanTissue) clock, 4% for the Horvath2 (Skin&Blood) clock, and 13% for the PhenoAge clock (Figure 4G). After PC-adjustment based on 50 , the technical variation between samples was reduced to 2%, 8%, and 2%, respectively (Figure 4G), indicating moderately robust technical validity at the single-CpG level, but high validity when multiple CpGs are combined into multivariate DNAm clocks. The DNAm dataset also contains 3 replicate longitudinal experiments of HC1 & HC2 that can be used to quantify the variability in the longitudinal rates of epigenetic aging using the investigator's preferred clock(s) or individual CpGs.…”