Patterns of epigenetic regulation of gene expression (and thus RNA and protein levels) change constantly in response to cell state and environment. Some of those changes are characteristic responses to the damage and dysfunction of aging. Since the demonstration of the first epigenetic clocks, those that predict age based on an algorithmic combination of the status of DNA methylation at CpG sites on the genome, researchers have produced any number of new clocks based on mining epigenomic, transcriptomic, proteomic, and other databases for correlations with age. Today's open access paper is yet another example of a new transcriptomic clock.
It remains the case that in none of these clocks is there is a good, well understood connection between specific mechanisms of aging and specific components of the clock algorithm. This makes it hard to make good use of aging clocks: it isn't at all clear that any given result is meaningful. If one applies a potentially rejuvenating or age-slowing intervention, and it produces a change in the clock measurements taken before and afterward treatment, what does that change mean? Is a drop in measured age a sign that the therapy is great, or a sign that the clock is overly weighted towards the subset of mechanisms of aging that are targeted by the intervention? If the clock shows little to no change, does that mean the therapy is useless, or the clock is unhelpful for this class of intervention? And so forth.
Thus clocks and therapies will have to be calibrated against one another in order to make the clocks useful. This process is only in the earliest stages, where it is occurring at all. As matters progress, this calibration will most likely mean running the slow, costly life span studies that we'd all like to avoid by using the clocks instead. There is no free lunch here.
Aging biomarkers that predict the biological age of an organism are important for identifying genetic and environmental factors that influence the aging process and for accelerating studies examining potential rejuvenating treatments. Diverse studies tried to identify biomarkers and predict the age of individuals, ranging from proteomics, transcriptomics, the microbiome, frailty index assessments to neuroimaging, and DNA methylation. Currently, the most common predictors are based on DNA methylation. The DNA methylation marks themselves might influence the transcriptional response, but aging also affects the transcriptional network by altering the histone abundance, histone modifications, and the 3D organization of chromatin. The difference in RNA molecule abundance, thereby, integrates a variety of regulation and influences resulting in a notable gene expression change during the lifespan of an organism. These changes sparked interest in the identification of transcriptomic aging biomarkers, an RNA expression signature for age classification, and the development of transcriptomic aging clocks.
While a large variety of data, techniques, and analyses have been used to identify aging biomarkers and aging clocks in humans, issues remain with regard to pronounced variability and difficulties in replicability. Indeed, a recent analysis of gene expression, plasma protein, blood metabolite, blood cytokine, microbiome, and clinical marker data showed that individual age slopes diverged among the participants over the longitudinal measurement time and subsequently that individuals have different molecular aging pattern, called ageotypes. These interindividual differences show that it is still difficult to pinpoint biomarkers for aging in humans.
Model organisms, instead, can give a more controllable view on the aging process and biomarker discovery. Caenorhabditis elegans has revolutionized the aging field and has vast advantages as a model organism. To date, no aging clock for C. elegans has been built solely on RNA-seq data and been shown to predict the biological age of diverse strains, treatments, and conditions to a high accuracy. In this study, we build such a transcriptomic aging clock that predicts the biological age of C. elegans based on high-throughput gene expression data to an unprecedented accuracy. We combine a temporal rescaling approach, to make samples of diverse lifespans comparable, with a novel binarization approach, which overcomes current limitations in the prediction of the biological age. Moreover, we show that the model accurately predicts the effects of several lifespan-affecting factors such as insulin-like signaling, a dysregulated miRNA regulation, the effect of an epigenetic mark, translational efficiency, dietary restriction, heat stress, pathogen exposure, the diet-, and dosage-dependent effects of drugs.
This combination of rescaling and binarization of gene expression data therefore allows for the first time to build an accurate aging clock that predicts the biological age regardless of the genotype or treatment. Lastly, we show how our binarized transcriptomic aging (BiT age) clock model has the potential to improve the prediction of the transcriptomic age of humans and might therefore be universally applicable to assess biological age.