Authors: Seungwon Kim*, University of Iowa
Topics: Medical and Health Geography, Spatial Analysis & Modeling, Quantitative Methods
Keywords: Network analysis, Phylogeography, Influenza A, Disease spread
Session Type: Paper
Start / End Time: 9:55 AM / 11:35 AM
Room: Tyler, Marriott, Mezzanine Level
Presentation File: No File Uploaded
Network models for the spread of infectious diseases have enhanced our understanding of underlying processes of disease transmission over space and time. However, previous epidemic models have critical limitations: 1) they assume fully mixed and fixed population structures which are unrealistic to represent the real processes of disease spread, and 2) they rely on the number of incidences within populations to define the edges. Increased availability of high resolution molecular and geographic datasets allows us to address these issues by incorporating a phylogeographic approach. Phylogeographic methods do not require the pre-defined population structures, rather they use molecular sequences to identify gene flows by comparing similarity/dissimilarity of gene sequences among different locations. For this research, gene sequences of influenza A in the US were collected from GenBank and the gene flows of influenza A virus among US states were inferred using BSSVS method. Two pairs of nodes were considered to be connected by a directed edge in the network topology if there was a significant flow as determined by the Bayes Factor. Then, a latent space model under a logistic regression approach was incorporated to address the dependency problem among edges in the network topologies and to evaluate the effect of climatic and demographic characteristics, and human transportation on disease spread. Results suggest that the network connectivity of influenza A viral gene flow is strongly influenced by climatic variables such as shared average humidity and temperature as well as by the degree of connectivity in the airline networks between US states.