Biological Sciences, School of


First Advisor

Etsuko N. Moriyama

Second Advisor

Luwen Zhang

Third Advisor

Chi Zhang

Date of this Version

Summer 7-25-2023


American Psychological Association


A THESIS Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Master of Science, Major: Biological Sciences, Under the Supervision of Professor Etsuko N. Moriyama. Lincoln, Nebraska: June 2023


The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of the current pandemic, Coronavirus Disease 2019 (COVID-19). SARS-CoV-2 is considered to be of zoonotic origin; it originated in non-human animals and was transmitted to humans. Since the early stage of the pandemic, however, the evidence of transmissions from humans to animals (reverse zoonoses) has been found in multiple animal species including mink, white-tailed deer, and pet and zoo animals. Furthermore, secondary zoonotic events of SARS-CoV-2, transmissions from animals to humans, have been also reported. It is suggested that non-human hosts can act as SARS-CoV-2 reservoirs where accumulated mutations in viral proteins could change the transmissibility and/or pathogenicity of the virus when it is spilled over again to human populations. Our goal, therefore, is to examine the SARS-CoV-2 genomic changes in non-human hosts and to identify the changes responsible for the adaptation of the virus in non-human hosts. Changes in the physicochemical properties of viral proteins potentially affect and influence their functions. Therefore, in this study, we compared SARS-CoV-2 proteins among human and non-human hosts and analyzed the differences in their physicochemical properties using the principal component analysis. In addition to the viral proteins from bat and pangolin, those from white-tailed deer and mink showed larger differences in the properties. Van der Waals volume, isoelectric point, charge, and thermostability index were found to be the main contributing factors. We next performed the comparisons of protein-protein interaction (PPI) prediction methods that use different features including physicochemical properties and those based on natural language processing. It showed that the Cross-attention PHV had slightly better performance scores than InterSPPI-HVPPI and LGCA-VHPPI. Finally, to examine the effect of changes in physicochemical properties in viral proteins against host proteins, PPI prediction was performed using the Cross-attention PHV between viral proteins from different SARS-CoV-2 variants and host proteins. The prediction scores between the different variants and host proteins from human and white-tailed deer were highly similar. The results showed that the analysis of physicochemical properties of viral proteins helps to understand how physicochemical properties of viral proteins affect viral-host PPIs and how viral proteins evolve to adapt different host cell environments.