The following scripts are contained within this repository and provide the analysis workflows that lead to data and figures of the publication "Cardiovascular disease biomarkers derived from circulating cell-free DNA methylation".

Raw tabix (and bigwig) files with associated metadata are publically available for discovery and validation cohort and can be downloaded via the links below.

The main content of each script is stated below:

• Alignment, coverage and methylation statistics discovery cohort
• Tiling of discovery cohort
• Comparative differential methylation (DM) analysis with healthy controls of discovery cohort
• GREAT gene set enrichment analysis for discovery differentially methylated regions (DMRs)

• Prediction of cell/tissue type proportions
• Plotting of Figure 2

• Annotation of DMRs to CpG islands, CpG shores, CpG shelves, and inter-CGI
• Annotation of DMRs to 1-5Kb, promoters, 5'UTRs, exons, introns, and 3'UTRs
• DisGNet enrichment analysis for discovery DMRs

• Design of probes for targeted sequencing of validation cohort using methylation atlas

• Assign targeted sequenced results for validation cohort to coordinates
• Get statistics for validation cohort

• Visualization and removal of sequencing bias between WGBS and targeted sequencing
• Validation of DMRS distinguishing between disease conditions (identified from discovery cohort)
• Plotting of Figure 6 and Suplementary Figure S6

• Correlation analysis clinical markers
• Distribution clinical markers per disease condition
• Bisulfite conversion rates
• Average CpG coverage
• Amount ccfDNA analysis
• Identification of 254 DMRs associated with disease groups on linear models
• Plotting Figure 1, 3, 4, 5 and Supplementary Figure S1 and S2

• builds upon ranking_v3.Rmd
• DMR analysis comparing each disease group vs. healthy controls
• Plotting of Supplementary Figure S4

• annotates DMRs for the groups, adding gene info, GO, phenotype associations, DO and disgenet associations

• Enrichment analysis, Fisher's exact test for cell types on DMRs
• Plotting of Figure S5

• Transfer DMRs distinguishing between disease conditions from discovery to validation cohort
• Plotting of Supplementary Figure S3

• Build predictive models to test multivariate predictive power of the 193 DMRs shared from discovery analysis and validation amples.

• evaluate the performance of our extended reference atlas/ CpG signature matrix compared to the comprehensive reference atlas generated by Moss et al (2018).

Download links for raw methylKit objects (tabix format) are provided for the discovery and validation cohort. Corresponding patient metadata are provided as tsv files for discovery and validation cohorts. Links to BigWig files are available for the discovery cohort.