Assessment of SARS-CoV-2 lineage alterations and regions of evolutionary variability to detect antigenically altered potential Variants of Interest

SARS-CoV-2 has a demonstrated capacity for rapid evolution through its accumulation of amino acid changes across the spike protein that improve viral fitness, particularly its ability to evade host immunity. Due to the global mass sequencing efforts throughout the COVID-19 pandemic, SARS-CoV-2 offers a unique opportunity to study the role of critical amino acid changes on viral antigenicity to optimize both public health management and vaccine efficacy. In this thesis I will delve into the prediction of SARS-CoV-2 antigenicity based on amino acid changes occurring throughout the spike protein of the virus. First, I identified patches of convergent evolution across the spike protein based on the recurrent substitution counts occurring at these clustered sites in an inferred phylogenetic tree. Key changes at these sites have demonstrated physicochemical property alterations that aid in the binding of the virus to host cells and improve its overall capacity for immune escape. Second, I developed a novel technique that scores SARS-CoV-2 lineages’ potential antigenicity and identifies significantly antigenically altered circulating lineages. This method successfully identified the Omicron variant and key sublineages known for reduced antibody neutralization and scored lineages significantly closer to known experimental fold reduction in neutralization scores compared to other scoring techniques such as SpikePro and EVEscape. The standardized antigenic alteration scores were also significantly higher for WHO-designated lineages compared to non-designated lineages with increasing order from Variants Under Monitoring (VUMs), to Variants of Interest (VOIs) to Variants of Concern (VOCs). Finally, to support ongoing genomic surveillance efforts, we developed CoVerage (www.sarscoverage.org), an online web-based platform that applies the scoring methodology alongside two additional methods to detect lineages and alleles on the spike protein with significant increases in frequency, termed potential Variants of Interest (pVOIs). CoVerage predicts and characterizes these emerging pVOIs with a lead time of almost three months relative to the lineages’ first official WHO designation and can be used for the assessment of future SARS-CoV-2 variants relevant for public health. Ultimately, the methods and resources I discuss in this thesis can be used to aid public health officials in the identification of antigenically novel SARS-CoV-2 lineages to better inform health policy and vaccine design.