Demystifying common DNA methylation sites that promote the ability of CheekAge to associate with health and disease

We recently showed that the next-generation epigenetic aging clock CheekAge was significantly associated with 33 different health and disease signals across 25 publicly available MethylationEPIC datasets. We additionally uncovered DNA methylation sites that played a disproportionately important role in driving the ability of CheekAge to associate with each of these variables. We dubbed these "pro" CpGs because of their ability to promote a given association. Here, we identify 2639 common DNA methylation sites that were a "pro" CpG for at least two different health and disease signals. Using genes annotated to these common "pro" CpGs, we perform extensive enrichment analyses to unveil motifs of DNA repair, cell division, tumors, cancer, and inherited syndromes. We additionally show that a plethora of genes linked to these common "pro" CpGs have been reported to alter lifespan and/or healthspan in model organisms when manipulated. Some of these were highly germane to canonical signaling pathways, such as RPTOR (encoding for Regulatory-associated protein of mTOR) and FOXA2 (encoding for Hepatocyte nuclear factor 3-beta). We also honed in on the seven DNA methylation sites that represented the most common "pro" CpGs, which were cg00005888, cg02478836, cg18331022, cg01892528, cg27634071, cg09232037, and cg17841124. Not only do we summarize available literature for these sites and their annotated genes, but we show that they can be combined into a proof-of-concept epigenetic biomarker that associates with alcohol intake, diet quality, and sex. All together, we provide additional insights into DNA methylation sites that underlie CheekAge's ability to associate with meaningful signals.