Fig. 1

High-level illustration of the computational framework to identify CpG sites causal to a phenotypic trait, and to elucidate their functionalities. Epigenome-Wide Mendelian Randomization (EWMR) is run to identify CpG sites causal to facial aging utilizing methylation quantitative loci (meQTL) from whole blood as exposure (see “Methods”). Summary statistics for \(\approx 11.2\) million SNP-CpG pairs reaching genome-wide significance are available [19]. For outcomes, summary statistics from genome-wide association study (GWAS) on perceived facial aging (ukb-b-2148) were obtained from the MR-Base GWAS catalog [20]. CpGs causal to facial aging were further interrogated by their overlap with epigenetic biological clocks available in the methylCIPHER R library. Functional enrichment analyses on the level of CpG were conducted using the methylgsa R library. Functional enrichment analyses on the level of nearby genes and associated proteins, obtained via the EWAS catalog, were conducted using the FUMA Functional Mapping tool.