Drivers genes in severe forms of COVID-19

The etiology of severe forms of COVID19, especially in young patients, remains a salient unanswered question. Here we build on a 3-tier cohort where all individuals/patients were strictly below 50 years of age and where a number of comorbidities were excluded at study onset. Besides healthy controls (N=22), these include patients in the intensive care unit with Acute Respiratory Distress Syndrome (ARDS) (“critical group”; N=47), and those in a non-critical care ward under supplemental oxygen (“non-critical group”, N=25). This highly curated cohort allowed us to perform a deep multi-omics approach which included whole genome sequencing, whole blood RNA-sequencing, plasma and peripheral-blood mononuclear cells proteomics, multiplex cytokine profiling, mass-cytometry-based immune cell profiling in conjunction with viral parameters i.e. anti-SARS-Cov-2 neutralizing antibodies and multi-target antiviral serology. Critical patients were characterized by an exacerbated inflammatory state, perturbed lymphoid and myeloid cell compartments, signatures of dysregulated blood coagulation and active regulation of viral entry into the cells. A unique gene signature that differentiates critical from non-critical patients was identified by an ensemble machine learning, deep learning and quantum computing approach. Within this gene network, Structural Causal Modeling identified several ARDS driver genes, among which the up-regulated metalloprotease ADAM9 seems to be a key driver. Inhibition of ADAM9 ex vivo interfered with SARS-Cov-2 uptake and replication in human epithelial cells. Hence we apply a machine learning approach to identify driver genes for severe forms of COVID-19 in a small, uncluttered cohort of patients.