The Initial COVID-19 Reliable Interactive DNA Methylation Markers and Biological Implications.

Simple Summary: A fundamental scientific question—where did SARS-CoV-2 come from?—has eluded the scientific community since it was first identified in December 2019. SARS-CoV-2/COVID-19 is still infecting humans. Like many other viruses, SARS-CoV-2 has been regarded as an RNA virus. However, the pathological knowledge of the cause of COVID-19 and the intrinsic drivers of virus replications are unknown. Finding an answer can help in understanding the virus and preventing the next pandemic. Many COVID-19 research results at the genomic level have been published in the literature. These published results have explored the pathological causes of COVID-19 infection from various aspects. Due to the limitations of research methodology, some published results can hardly be cross-validated from cohort to cohort. As a result, there is still an urgent need to study the root causes further. We aim to find an answer to the fundamental scientific question. We use a new AI algorithm to identify critical genes at the DNA methylation level. Our results are computational with biological implications. They are interpretable, accurate, and cross-validated. They can be reproduced from Excel sheets using the derived formula. Our findings demand rigorous and much deeper study. Earlier research has established the existence of reliable interactive genomic biomarkers. However, reliable DNA methylation biomarkers, not to mention interactivity, have yet to be identified at the epigenetic level. This study, drawing from 865,859 methylation sites, discovered two miniature sets of Infinium MethylationEPIC sites, each having eight CpG sites (genes) to interact with each other and disease subtypes. They led to the nearly perfect (96.87–100% accuracy) prediction of COVID-19 patients from patients with other diseases or healthy controls. These CpG sites can jointly explain some post-COVID-19-related conditions. These CpG sites and the optimally performing genomic biomarkers reported in the literature become potential druggable targets. Among these CpG sites, cg16785077 (gene MX1), cg25932713 (gene PARP9), and cg22930808 (gene PARP9) at DNA methylation levels indicate that the initial SARS-CoV-2 virus may be better treated as a transcribed viral DNA into RNA virus, i.e., not as an RNA virus that has concerned scientists in the field. Such a discovery can significantly change the scientific thinking and knowledge of viruses. [ABSTRACT FROM AUTHOR]