Ancestry-specific DNA methylation signatures of smoking and associations with atherosclerotic cardiovascular disease risk: findings from the Million Veteran Program

DNA methylation is a key epigenetic biomarker linking environmental exposures to downstream effects on health and disease, and tobacco smoking is among the strongest known influences on the epigenome. However, even the largest previous meta-analyses of smoking have been limited by relatively low diversity of study populations and sample size, and no previous studies have been able to examine well-powered ancestry-specific epigenetic signals of smoking. We performed epigenome-wide association analyses of tobacco smoking in a large, diverse sample of 42,450 participants from the Million Veteran Program, a nationwide cohort of US Veterans. We observed 40,529 and 4,240 Bonferroni-significant sites associated with current and former smoking, respectively, in the overall population. Furthermore, we observed substantial commonality in the associations identified across ancestry-specific analyses (conducted within European, African, and Admixed American populations), with 95–100% of ancestry-specific associations also identified in overall population and 31 CpG sites differentially methylated across all analyses, as well as evidence of ancestry-specific associations and differences in association magnitudes between ancestry groups. A methylation risk score generated from these 31 common CpG sites modestly improved the prediction of time to atherosclerotic cardiovascular disease (concordance = 0.673) relative to three-category smoking status (concordance = 0.668). We performed the largest epigenome-wide association study of smoking activity to date, reporting extensive validation of previous findings, novel findings and ancestry-specific associations, and extensive low magnitude associations related to former smoking, further finding that smoking-associated changes in DNA methylation improve the prediction of atherosclerotic cardiovascular disease.