Genomic Signatures of SARS-CoV-2 Associated with Patient Mortality

Infections with SARS-CoV-2 can progress toward multiple clinical outcomes, and the identification of factors associated with disease severity would represent a major advance to guide care and improve prognosis. We tested for associations between SARS-CoV-2 genomic variants from an international cohort of 2508 patients and mortality rates. Findings were validated in a second cohort. Phylogenetic analysis of SARS-CoV-2 genome sequences revealed four well-resolved clades which had significantly different mortality rates, even after adjusting for patient demographic and geographic characteristics. We further identified ten single-nucleotide polymorphisms (SNPs) in the SARS-CoV-2 genome that were associated with patient mortality. Three SNPs remained associated with mortality in a generalized linear model (GLM) that also included patient age, sex, geographic region, and month of sample collection. Multiple SNPs were confirmed in the validation cohort. These SNPs represent targets to assess the mechanisms underlying COVID-19 disease severity and warrant straightforward validation in functional studies.