A Computational study of Ebolavirus Pathogenicity and a Modeling approach for human non-synonymous variants

Recent advances in genome sequencing are improving our better understanding of

genetic variation. However, the investigation of the genotype-phenotype relationship

is still challenging, especially for the interpretation of the myriad of discovered

genetic variants that weakly relate to disease.

Recently, researchers have confirmed that disease causing genetic variants typically

occur at functional sites, such as protein-protein or protein-ligand interaction sites.

Giving this observation, several bioinformatics tools have been developed. This

thesis first details VarMod (Variant Modeller), an algorithm that predicts whether

nonsynonymous single nucleotide variants (nsSNVs) affect protein function.

The recent Ebola virus outbreak in West Africa demonstrated the potential for the

virus to cause edipdemics and highlighted our limited understanding of Ebola virus

biology. The second part of this thesis focuses on the investigation of the molecular

determinants of Ebolavirus pathogenicity. In two related analyses knowledge of

differing pathogenicity of Ebolavirus species is used. Firstly, comparison of the

sequences of Reston viruses (the only Ebolavirus species that is not pathogenic in

humans) with the four pathogenic Ebolavirus species, enabled the identification of

Specificity Determining Positions (SDPs) that are differentially conserved between

these two groups. These SDPs were further investigated using analysis of protein

structure and identified variation in the Ebola virus VP24 as likely to have a role in

determining species-specific pathogenicity. The second approach investigated

rodent-adapted Ebola virus. Ebola virus is not pathogenic in rodents but it can be

passaged to induce pathogenicity. Analysis of the mutations identified in four

adaption studies identified that very few mutations are required for adaptation to a

new species and once again the VP24 is likely to have a central role. Subsequent

molecular dynamics simulations compared the interaction of Ebola and Reston virus

VP24 with human karyopherin alpha5. The analysis suggests that Reston virus VP24

has weaker binding with karyopherins and we propose that this change in binding

may reduce the ability of Reston VP24 to inhibit human interferon signaling.