Study design for epigenome-wide analyses on prevalent and incident disease states in Generation Scotland.

(A) Recruitment for Generation Scotland took place between 2006 and 2011. Prevalence analyses: participants self-reported disease status and donated blood samples at the study baseline. Incidence analyses: linked healthcare data were used to determine if participants who were free from a particular condition at baseline went on to develop the condition over up to 14 years of follow-up. Controls were free of the disease at the baseline and during follow-up. (B). Blood DNAm at baseline was available for 18,413 participants. The mean age was 47.5 years and the sample was 58.8% female. EWAS tested for associations between blood CpG methylation and the prevalence of 14 disease states at baseline or the incidence (time-to-onset) of 19 disease states during follow-up. The mean time-to-diagnosis ranged from 5.0 years (for chronic pain) to 11.7 years (for COVID-19 hospitalisation). Significant findings were tested for replication in existing studies via a structured literature review. Replication within existing studies was also investigated. Colocalisation analyses were employed to help dissect whether associations between DNAm and disease states reflected shared or distinct genetic architectures. (C). The first box lists the 14 self-reported disease states at the study baseline, which were included in this study. The second box lists the 19 incident disease states identified through electronic health record linkage. They include the same 14 conditions listed in the first box along with 5 additional disease states. Of note, prevalent AD reflected family history of the disease due to the young mean age of the sample at baseline, whereas incident AD reflected diagnosed disease. Image created using Biorender.com. AD, Alzheimer’s dementia; COVID-19, Coronavirus Disease 2019; CpG, cytosine-phosphate-guanine dinucleotide; DNAm, DNA methylation; EWAS, epigenome-wide association studies.