Abstract
Investigating the role of DNA methylation in the development of pancreatic cancer (PC) may facilitate identification of potential targets for both diagnosis and treatment. We carried out a comprehensive epigenome-wide Mendelian randomization (EWMR) analysis to investigate the correlation of genetically predicted blood CpG sites with PC. Following this, we conducted various sensitivity analyses and repeated analyses using different selection criteria for instrumental variables and conditional Bayesian colocalization to guarantee the reliability of the results. External validation and a meta-analysis were then performed to further validate these results. Next, we conducted CpG site enrichment analysis, overlap with phenome-wide association studies (PheWAS) catalog analysis, overlap with epigenome-wide association studies (EWAS) Toolkit analysis, and drug target analysis to explore the enrichment, biological functions, and potential therapeutic targets associated with these sites. Finally, we used the SMR-IVW software to perform mediation analysis, aiming to uncover potential tumorigenesis pathways of PC at the transcriptional level from three distinct perspectives. Results showed 253 CpG sites passing sensitivity analysis were significantly associated with PC and 159 CpG sites were validated in at least one replication. After meta-analysis, 38 CpG sites were retained, and all 253 CpG sites were classified into three tiers. Among these, cg26373071 (CLPTM1L), cg14271713, cg11652496 (PSTPIP1), and cg20575191 (PSTPIP1) were placed in tier 1 with strong support. Finally, this study identified genetic susceptibility linked to 253 PC-related CpG sites. This study provides insights into the disease's origins and underscores potential targets for future research.