2.2. DNA Methylation Data

DNA was extracted from stored blood buffy coats using the QIAamp DNA Blood Kit (Qiagen, Valencia, CA, USA). We treated 500 ng DNA for bisulfite conversion using the EZ-96 DNA Methylation Kit (Zymo Research, Orange, CA, USA). After bisulfite conversion, we hybridized DNA samples to 12-sample Illumina HumanMethylation450 BeadChips per Infinium HD Methylation protocol (Illumina, San Diego, CA, USA). Technical effects resulting from plates/chips were minimized with a two-stage, age-stratified algorithm to randomize the samples. The β-values, ranging from 0–1, were used to represent the methylation status of a specific CpG site, with 0 indicating no methylation and 1 indicating full methylation.

Calculation of DNAmPhenoAge: DNAmPhenoAge was estimated using the method described by Levine et al. [10]. In short, phenotypic age was developed using clinical measurements from the third National Health and Nutrition Examination Survey. A penalized Cox regression model was used to select 10 out of 42 biomarkers. Phenotypic age was then regressed on DNA methylation data from whole blood samples based on an elastic net regression analysis, which led to 513 CpG sites. We calculated DNAmPhenoAge based on these loci for each participant as:

Values of coefficients and intercepts were extracted from Levine et al. [10]. DNAmPhenoAge acceleration was defined as residuals from a linear regression of DNAmPhenoAge on chronological age.

Selection of smoking-related DNA methylation loci: We selected 151 CpG sites reported from previous smoking EWAS at least twice from a previous systematic review [3]. There was no overlap between these 151 loci and the 513 loci used to predict DNAmPhenoAge.