Divergent causal signatures of the adiposity quadrants on coronary artery disease: a two-sample Mendelian randomization and mediation study — data and code

This deposit contains the summary-level data, derived genetic instruments, analytical results, and code supporting a two-sample Mendelian randomization (MR) study of the four quadrants of the adiposity–metabolic health plane (metabolically healthy or unhealthy × obese or lean) and their causal effects on coronary artery disease (CAD). The study tests the hypothesis that cardiovascular risk follows adipose storage dysfunction rather than body weight per se. Each quadrant is genetically instrumented; total causal effects on CAD are estimated by inverse-variance weighted MR with a full sensitivity framework (MR-Egger, weighted median, MR-PRESSO, Steiger, heterogeneity); and the mediating conduits are dissected using a harmonized 29-mediator panel and multivariable MR (MVMR). The unhealthy-lean insulin-resistance axis is independently replicated with a second, methodologically distinct instrument (mvIR, 2025) and across three CAD cohorts (CARDIoGRAMplusC4D, Hartiala 2021, FinnGen). The deposit includes eight result tables, six figures, and the analytical R code. All primary GWAS sources are publicly available and cited in the README.