The founder of the PhenoAge epigenetic clock methodology authored this 2020 article:
“The Ge[r]oscience paradigm suggests that targeting the aging process could delay or prevent the risk of multiple major age-related diseases. We need clinically valid measures of the underlying biological process and/or classification criteria for what it means to be biologically, rather than chronologically, “aged”.
The majority of aging biomarkers, including the first-generation epigenetic clocks, are developed using cross-sectional data, in which the researchers take a variable that proxies aging (e.g. chronological age) and apply supervised machine learning, or deep learning, approaches to predict that variable using tens to hundreds of thousands of input variables. The problem with this approach is that it doesn’t account for mortality selection. This biases the algorithm to select markers that are not causal, but instead correlative with aging.
When considering individuals of the same chronological age, do those with higher epigenetic age look phenotypically older on average (e.g. have higher mortality rates, greater disease burden, and worse physical and cognitive functioning)? FEV1 [forced expiratory volume in one second] declined at a faster rate for individuals with higher baseline GrimAge and/or PhenoAge. A similar finding was observed for the decline in grip strength as a function of GrimAge; however, the rate of change for any of the epigenetic clocks was not associated with rate of change in any performance measure.
Loci that show consistent trends with chronological age, even at higher ages, are likely not causal. By using a cross-sectional study design for biomarker development there was a propensity away from selecting causal loci, to the point where fewer causal loci were selected than if loci had been chosen at random.
The power of these measures as diagnostic and prognostic may stem from the use of longitudinal data in training them. Rather than continuing to train chronological age predictors using diverse data, it may be more advantageous to retrain some of the existing measures by predicting longitudinal outcomes.”
https://academic.oup.com/biomedgerontology/advance-article-abstract/doi/10.1093/gerona/glaa021/5717592 “Assessment of Epigenetic Clocks as Biomarkers of Aging in Basic and Population Research” (not freely available)
A cited 2019 study modeled corrections to “account for mortality selection.” It modified datasets “by incorporating correlates of mortality identified from longitudinal studies, allowing cross-sectional studies to effectively identify the causal factors of aging.”
https://academic.oup.com/biomedgerontology/advance-article-abstract/doi/10.1093/gerona/glz174/5540066 “Biomarkers for Aging Identified in Cross-sectional Studies Tend to Be Non-causative” (not freely available)
The article didn’t present a complete case to determine whether an individual’s epigenetic clock measurements over time may show causes of biological aging.
Other viewpoints include:
1. A blood plasma aging clock presented evidence with its 46-protein conserved aging signature that some causes of biological aging are under genetic control. If the principle of this finding applies to CpG DNA methylation, the statement:
Loci that show consistent trends with chronological age, even at higher ages, are likely not causal.
may not hold. Such epigenetic changes could be among both the causes of senescence and the effects of evolution’s selection mechanisms.
2. An epigenetic clock review by committee particularly in:
- Challenge 3 “Integration of epigenetics into large and diverse longitudinal population studies”;
- Challenge 5 “Single-cell analysis of aging changes and disease”; and
- Table 1 “Major biological and analytic issues with epigenetic DNA methylation clocks” with single-cell analysis as the solution to five Significant issues.