Any sufficiently large database of biological data obtained from individuals of different ages can be used to build an aging clock. Machine learning approaches can be used find algorithmic combinations of measures that, on average, map to chronological age, mortality risk, or a similar benchmark in the study population. Individuals with a clock age higher than their chronological age are then said to exhibit accelerated aging. The degree to which this approach actually produces good measures of biological age, meaning the burden of damage and dysfunction that causes eventual mortality, remains open to debate. It seems clear from the work conducted to date that some useful form of consensus assessment of aging will emerge eventually, and be used to speed up the development of therapies capable of reducing that measure of biological age.
New aging clocks are produced at a fair pace, a dozen or more every year at this point, even as many researchers are pushing for more of a focus on just a few specific clocks, trying to forge some consensus for a universally agreed upon clock to assess biological age. Today’s open access paper is an example of yet another new clock. Here, researchers expand on recently published work to describe their metabolomic clock called MileAge, built on metabolite levels derived from blood samples in a human study population.
The increasing availability of high-dimensional molecular omics and neuroimaging data, for example, DNA methylation (DNAm) and magnetic resonance imaging, has enabled the development of biological aging clocks. These clocks are typically developed using statistical or machine learning algorithms that identify relationships between chronological age and molecular data. The difference between predicted age and chronological age can track with health outcomes. Aging clocks provide a more holistic view of a person’s health and are more readily interpretable than many individual molecular markers, as they are expressed in units of years.
Metabolomics, the study of small molecules within cells, tissues, or organisms, is increasingly incorporated into biological aging research. Metabolites are the end products of metabolism, such as when food is converted to energy. Early metabolomics studies were limited to a few metabolites and small samples, but technological advancements have enabled the population-scale profiling of multiple molecular pathways. Quantifying hundreds or thousands of metabolites can provide detailed snapshots of an individual’s physiological state. Metabolomic profiles can predict many common incident diseases and mortality risk.
This study aimed to benchmark machine learning algorithms for developing metabolomic aging clocks from nuclear magnetic resonance spectroscopy data. The UK Biobank data, including 168 plasma metabolites from up to N = 225,212 middle-aged and older adults (mean age, 56.97 years), were used to train and internally validate 17 algorithms. Metabolomic age (MileAge) delta, the difference between metabolite-predicted and chronological age, showed the strongest associations with health and aging markers. Individuals with an older MileAge were frailer, had shorter telomeres, were more likely to suffer from chronic illness, rated their health worse, and had a higher all-cause mortality hazard (hazard ratio = 1.51). MileAge can be applied in research and may find use in health assessments, risk stratification, and proactive health tracking.
Trending Products
