With a case fatality rate higher than the 1918 Spanish Flu pandemic, H5N1 highly pathogenic avian influenza represents a threat to global public health. Efforts to identify locations with the greatest potential for pandemic emergence, as well as how the virus is spreading, may help minimize this threat. First detected in Egypt in 2006, H5N1 viruses have resulted in the deaths of millions of birds in both commercial and backyard poultry flocks, and more than 350 human infections, the most of any country, have been confirmed. Human outbreaks have been so far constrained by poor viral adaptation to non-avian hosts. There are two evolutionary mechanisms by which the H5N1 avian influenza virus could acquire pandemic potential: 1) via reassortment as a result of coinfection with another subtype (such as low pathogenic avian influenza H9N2); and/or 2) via antigenic drift and the accumulation of randomly occurring genetic changes found to improve viral fitness, herein called key substitutions (KS). Both mechanisms were investigated using geospatial methods including ecological niche modeling and hot spot analyses to predict locations with elevated potential for pandemic emergence. Using ecological niche modeling environmental, behavioral, and population characteristics of H5N1 and H9N2 niches within Egypt were identified, with niches differing markedly by subtype. Niche estimates were combined using raster overlay to estimate co-infection potential, with known occurrences used for validation. Co-infection was successfully predicted with high accuracy (area under the receiver operating characteristic (ROC) curve (AUC) 0.991). 41 distinct KS in H5N1 were detected in Egyptian isolates, including 17 not previously reported in Egypt. Phenotypic consequences of detected KS were varied, but the majority have been implicated in improving mammalian host adaptation and increasing virulence. Statistically significant spatial clustering of high KS rates was detected in the northwestern portion of the Nile River delta in the governorates of Alexandria and Beheira. To investigate how the virus spreads between poultry farms, landscape genetics techniques were employed. Viral genetic sequences were evaluated using phylogenetics to determine viral relatedness between samples, then distance models representing competing diffusion mechanisms were created using road networks and a least-cost path model designed to approximate wild waterbird travel using niche modeling and circuit theory. Spatial correlations were evaluated using Mantel tests, Mantel correlograms, and multiple regression of distance matrices within causal modeling and relative support frameworks. Samples from backyard farms were most strongly correlated with least cost path distances, implicating wild bird diffusion, while samples from commercial farms were most strongly correlated with road network distances, implicating human-mediated diffusion. Results were largely consistent across gene segments. Identifying areas at risk of co-infection can help target spaces for increased surveillance. Similarly, detecting spatial hot spots of KS highlight areas of concern for pandemic emergence from antigenic drift. Demonstration of different diffusion mechanisms by farm type should inform both surveillance and biosecurity practices. Knowledge of where to focus intervention efforts, both spatially and strategically, allows limited public health resources to be targeted most effectively. By detecting where in the country pandemic influenza is likely to emerge and identifying how the virus is spreading between farms, this work contributes to efforts to predict and prevent the next influenza pandemic.
Identifer | oai:union.ndltd.org:uiowa.edu/oai:ir.uiowa.edu:etd-7167 |
Date | 01 May 2017 |
Creators | Young, Sean Gregory |
Contributors | Carrel, Margaret |
Publisher | University of Iowa |
Source Sets | University of Iowa |
Language | English |
Detected Language | English |
Type | dissertation |
Format | application/pdf |
Source | Theses and Dissertations |
Rights | Copyright © 2017 Sean Gregory Young |
Page generated in 0.0021 seconds