Search

As protein sequences evolve, differences in selective constraints may lead to outcomes ranging from sequence conservation to structural and functional divergence. Evolutionary protein family analysis can illuminate which protein regions are likely to diverge or remain conserved in sequence, structure, and function. Moreover, nonsynonymous mutations in pathogens may result in the emergence of protein regions that affect the behavior of pathogenic proteins within a host and host response. I aimed to gain insight on pathogenic proteins from cancer and viruses using an evolutionary perspective. First, I examined p53, a conformationally flexible, multifunctional protein mutated in ~50% of human cancers. Multifunctional proteins may experience rapid sequence divergence given trade-offs between functions, while proteins with important functions may be more constrained. How, then, does a protein like p53 evolve? I assessed the evolutionary dynamics of structural and regulatory properties in the p53 family, revealing paralog-specific patterns of functional divergence. I also studied flaviviruses, like Dengue and Zika virus, whose conformational flexibility contributes to antibody-dependent enhancement (ADE). ADE has long complicated vaccine development for these viruses, making antiviral drug development an attractive alternative. I identified fitness-critical sites conserved in sequence and structure in the proteome of flaviviruses with the potential to act as broadly neutralizing antiviral drug target sites. I later developed Epitopedia, a computational method for epitope-based prediction of molecular mimicry. Molecular mimicry occurs when regions of antigenic proteins resemble protein regions from the host or other pathogens, leading to antibody cross-reactivity at these sites which can result in autoimmunity or have a protective effect. I applied Epitopedia to the antigenic Spike protein from SARS-CoV-2, the causative agent of COVID-19. Molecular mimicry may explain the