Epigenetics of Covid-19: A Systems Biology Approach

This thesis investigates how DNA methylation serves as a critical regulatory layer that explains the extreme variability in COVID-19 outcomes, moving beyond the limitations of purely genetic explanations. By conducting an epigenome-wide association study on blood samples, the research initially identified 880 CpG sites significantly associated with disease severity, many of which are located within genes central to immune and inflammatory responses like SAMHD1 and NLRP3. While these individual probes highlighted biological processes such as interferon signaling and antiviral defense, they were initially insufficient for clear patient classification on their own. To gain a deeper understanding, the study adopted a systems biology approach using Weighted Gene Co-expression Network Analysis to identify modules of coordinated methylation. This revealed that COVID-19 severity is linked to broad shifts in immune, metabolic, and structural pathways, including B-cell signaling and viral entry mechanisms. From these networks, a refined signature of 55 CpGs was extracted using advanced statistical modeling. This signature proved highly effective at separating mild from severe cases across different patient cohorts, maintaining its predictive power regardless of the age or sex of the individual. The research culminated in the development of predictive machine learning models which were made transparent through the use of SHAP values. This allowed for both global insights into which methylation changes drive disease progression and local, patient-specific explanations of clinical risk. Despite limitations such as a modest sample size and the use of blood rather than lung tissue, the work demonstrates that DNA methylation encodes vital information about disease pathology. Ultimately, the thesis provides a reproducible framework for using epigenomics and machine learning to bridge the gap between genetic risk and clinical reality, offering potential new biomarkers for infectious and chronic diseases.