We hope to identify genomic markers that causally influence susceptibility to acquiring COVID-19, the severity and duration of symptoms of the disease, the length of time the virus can be detected, and the occurrence of death due to the disease. Knowing these regions can support (1) the identification of people at risk, (2) the triage of patients who become infected, and (3) improvements in treatment of the disease.
To identify these genomic markers, we will perform a genome-wide association analysis (GWAS). One common issue with GWAS is that associations found can be non-causal, and therefore of little use to guide interventions. In GWAS, non-causal associations often arise because genomic markers can be shared among extended families. In the case of GWAS for COVID-19, non-causal associations can also arise from the fact that the disease spreads in space and in time. In this work, we use a statistical model that can simultaneously account for shared genomics, geographical location of infection, and time of infection, and thus more likely to reveal true causal associations.