Date of Award

Fall 12-2018

Degree Name

Master of Public Health (MPH)



Second Advisor



Many studies of Respiratory Syncytial Virus (RSV) have relied on analyses of the Major Surface Glycoprotein G gene (G gene). Global transmission patterns have not been well studied due to lack of systematic global surveillance efforts. This study used phylogenetic analysis of full genome data, categorized by geo-region, to determine the sources of RSV A and B infection in Chile and Houston, Texas. Additionally, disease severity studies have generally focused on outcomes associated with a single genotype. In this study we developed a statistical phylogenetic approach to explore the relationship between tree topology and disease severity. Disease severity data included if the patient was given oxygen, if they were hospitalized, and if they were admitted to an intensive care unit.

Global data was downloaded from GenBank, separated into RSV A and RSV B, aligned, and manually optimized. The United States and Canada region was overrepresented in the publicly available data, so subsampling was conducted to reduce selection bias. Starting trees were generated from the subsampled datasets using RAxML. Geographic traits and trait state transition rates were jointly estimated in a Bayesian statistical framework using BEAST. The global transmission network was estimated using the Bayesian stochastic search variable selection and a constant population with a HKY genetic substitution model. Trait associations were calculated using BaTS. For RSV A, the time to most recent common ancestor (tMRCA) was 1963.40 (95% BCI: 1946.15, 1969.60). For RSV B, the tMRCA was 1963.80 (95% BCI: 1959.50, 1967.33). Europe and Central Asia was a key source of RSV A and B transmissions for both Chile and Houston. In addition, the Middle East and North Africa and Latin America and the Caribbean were sources of RSV transmission into Houston.

For the RSV A clinical data, there were significant associations between disease severity and tree topology when analyzing all three traits together (AI 3.13 p<0.01, PS 22.39 p<0.01) and for oxygen (AI 0.98 p<0.01, PS 9.32 p<0.01) and hospitalization independently (AI 1.92 p<0.01, PS 11.78 p<0.01). No significant association was found between tree topology and ICU admission. No significant associations were found in the RSV B clinical data, which may be due to the small sample size and homogeneous outcomes in this group.

Improved surveillance systems are needed to gain a better understanding of global transmission patterns to complement studies done of local transmission patterns, as global introductions play an important role in local outbreaks. Identifying genetic mutations that lead to more severe outcomes may help researchers target vaccine development.