Compatibility and comparative analysis of chronological and biological aging between the legacy 450K and the EPIC v2.0 arrays

Abstract

Several different epigenetic clocks built from DNA methylation beadchip arrays, including the Illumina Infinium Human Methylation 450K and EPICv1 BeadChips, have been developed and are used to evaluate biological aging. However, there is still limited information on effectively utilizing the novel EPICv2 arrays and addressing biases in epigenetic clock calculations resulting from missing probes initially present on legacy arrays. We address how the absence of probes on the EPICv2 array, originally present on the 450K array, affects many available epigenetic clock estimates. Using data from the InCHIANTI study of aging, our findings show that while clocks like Hannum and GrimAgeV2 exhibit significant over- or underestimation due to missing probes, the correlations between estimates from all 450K probes (unreduced) and those from shared probes between 450K and EPICv2 (reduced) remain strong. In this cohort, we consider AgeAcceleration from traditional epigenetic clocks to be the most reliable approach. Their values do not significantly change upon probe reduction, with unreduced and reduced values being nearly identical and with stable mean differences across chronological age.