Framingham Offspring Exam 8 and Generation 3 Exam 2 450K Array DNA Methylation Study

This substudy phs000724 Framingham DNA Methylation contains raw 450K offspring DNA methylation data as well as processed DNA methylation in both offspring and generation3 (Gen3) participants. 850K array DNA methylation data in both offspring and Gen3 can be found under the FHS substudy [phs000974](./study.cgi?study_id=phs000974) NHLBI TOPMed: The Framingham Heart Study.

Summary level phenotypes for the Framingham Cohort study participants can be viewed at the top-level study page [phs000007](./study.cgi?study_id=phs000007) Framingham Cohort. Individual level phenotype data and molecular data for all Framingham top-level study and substudies are available by requesting Authorized Access to the Framingham Cohort study [phs000007](./study.cgi?study_id=phs000007).

The risk of cardiovascular disease (CVD) is determined by the complex interaction of multiple genetic and environmental factors. Although genome-wide association studies (GWAS) have been highly successful in unraveling genes and pathways involved in multiple complex traits and diseases, common genetic variation only explains a small proportion of the heritability of these traits. It is believed that epigenetic factors (modifications in how our genes work that are not due to changes in DNA sequence) impact greatly on multiple complex traits. Epigenetic modifications, however, are largely unexplored in cardiovascular disease (CVD), where many of the causal genes are known to be regulated by epigenetic mechanisms, including DNA methylation. We seek to characterize DNA methylation in participants from the Framingham Heart Study in order to characterize the contribution of DNA methylation to CVD and other complex traits. This project, when combined with the vast data resources of the Framingham Heart Study (including gene expression and GWAS), will further the knowledge of CVD prevention, prediction, and therapy.

For Offspring participants, both raw and processed DNA methylation is available. Offspring DNA methylation from exam 8 was quantified by two labs: 576 samples (CVD case-control study samples) were tested in John Hopkins University and 2270 remaining offspring samples were tested in University of Minnesota.

For Gen3 participants, only processed DNA methylation is available. For Gen3 exam 2 sample selection, family study design was emphasized. We selected ~ 1,080 Gen3 participants with both parents having whole genome sequencing (WGS) and DNA methylation data. We also selected ~ 500 Gen3 participants with mother or father having WGS and DNA methylation data. All Gen3 participants were tested by the Illumina lab.

For the processed DNA methylation results, lab-specific normalization and quality control (QC) procedures were performed on all samples. QC procedures included multidimensional scaling (MDS) analysis to identify sex mismatch and outliers, call rate checking for samples and DNA probes, consistency checking by comparing SNPs for 65 overlap SNPs by previous genotyping/1000 Genome imputation and by Illumina 450K methylation array, and exclusion of probes that mapped to multiple locations and had SNPs at CpG sites or ≤10 bp of Single Base Extension. After QC procedures, we excluded samples that were MDS outliers, had high missing rate (>1%), or poor matching to SNP genotype. Probes were excluded if they had high missing rate (>20%), mapped to multiple locations, had SNP (MAF>5% in EUR 1000G) at CpG site or ≤10 bp of Single Base Extension.

After QC procedures, 2152 samples with 444,098 probes remained for the UMN-lab; 485 samples with 444,055 probes remained for the JHU-lab; 1522 samples with 444046 probes remained for the gen3-lab. The data is packaged separately for each of the labs; each have 444 ‘common' packages with 1000 probes each. The UMN, JHU and gen3 versions with the same package number contain the same DNA methylation probes in the same order. In addition to the 444 common packages, there are lab-specific packages for UMN only (n=98 probes), JHU only (n=55 probes), and gen3 only (n=44 probes). Please note, these 98, 55 and 44 probes are not in common in three labs. But there are overlapping probes between the 98 and 55, 98 and 44, and 55 and 44. Post QC, the total sample size is 4159.

Project aims:

1.\tTo assess differences in DNA methylation across the genome, in relation to cardiovascular disease and other phenotypes - leveraging the extensive surveillance for events and comprehensive phenotype characterization in Framingham.

2.\tTo characterize the relations of site specific DNA methylation to transcription levels across the genome at a gene and exon level -- leveraging the genomic resources of the SABRe CVD Initiative.

3.\tTo relate lifestyle and environmental exposures (e.g. diet, smoking, obesity, sleep) and other factors (e.g. sex hormone status) to DNA methylation by investigating the locations, levels and frequency of DNA methylation in relation to these traits -- leveraging the phenotypic resources of Framingham.

4.\tTo relate common genetic variation from GWAS to variable DNA methylation -- leveraging the genetic resources of the Framingham SHARe GWAS.