Additional file 1: of Analysis of repeated leukocyte DNA methylation assessments reveals persistent epigenetic alterations after an incident myocardial infarction

Supplemental methods, Table S1. CpGs from KORA EWAS. Those CpGs with a false discovery rate P < 0.15 in an epigenome-wide association study in KORA using the difference in methylation between the baseline and follow-up exams as the outcome (after adjustment for technical factors) and the occurrence of an MI as the predictor while adjusting for clinical covariates (at both baseline and follow-up) in a generalized estimating equations model. Table S2. The 174 CpGs which were retained from the initial elastic net model performed in KORA. Table S3. Epigenetic loci with non-zero coefficients from the NAS elastic net model. Table S4. AUC for the model fit with the loci with non-zero betas from the NAS elastic net in KORA, NAS, and InCHIANTI. Table S5. Medication usage in KORA at baseline and follow-up. We divide out the medication usage in KORA at (a) baseline and (b) follow-up for those individuals who did not develop and incident MI during the observation time (MI free) vs those that did (MI cases). Table S6. Association between 11 epigenetic fingerprint loci and medications. Associations were performed relative to both starting and stopping six classes of medications: diuretic, beta-blockers, anti-platelet, calcium channel blocker, statins, ACE-inhibitor, and angiotensin inhibitor. Table S7. Count of the genes within 1 Mb of each epigenetic fingerprint loci. Table S8. Integration of methylation, gene expression, and metabolomics for the suggestive (P < 0.001) gene expression-metabolite associations. Figure S1. Post hoc power estimations for the observed effects at our FDR cutoff of 0.15 for the initial screening EWAS. (ZIP 196 kb)