The ageing process involves complex biological changes, with epigenetic alterations playing a crucial part. Epigenetic clocks, the machine learning algorithms that predict biological age on the basis of DNA methylation patterns, have become popular computational tools in ageing research. These clocks can accurately estimate chronological age and have been associated with various health outcomes, including age-related diseases and mortality. However, these clocks are based on correlations, making it difficult to distinguish causal factors from mere associations in the ageing process.

In our recent study, we applied EWMR to investigate the causal relationship between DNA methylation and various ageing-related traits, including lifespan, healthspan and frailty. We discovered numerous CpG sites whose methylation levels have a causal effect on ageing-related traits. These sites provide potential targets for future interventions aimed at modulating the ageing process. Causal CpG sites identified through EWMR were found to be enriched in specific regulatory regions and transcription factor binding sites, providing insights into the molecular mechanisms underlying the epigenetic regulation of ageing. Interestingly, we found that existing epigenetic clocks are not significantly enriched for causal CpG sites identified by EWMR. This finding highlights the importance of causal inference to develop more accurate biomarkers of ageing.