Epidemic dynamics are key in basic and applied population biology, and the application of social network models to the spread of infectious diseases is an intuitive refinement to the classical assumption of population-wide random mixing. However, there is currently a disjunction between existing mathematical models for contact networks, which underlie epidemic dynamics, and real-world network data. The proposed project will address this gap by developing a new synthesis of recently developed statistical methods. In most systems, although observed data on the true contact network are unavailable, ancillary data (host and pathogen genetic data and coarse-grain data on social and spatial relationships) exist that provide information about the contact network. Yet there are formidable mathematical challenges in developing probability models for contact networks using data that do not include an actual observed network. Even if such a model were obtained, conducting predictive inference on epidemic dynamics requires a simulation framework that will generate unbiased realizations from the network model on which we can test epidemic and evolutionary processes. This project will develop methods to address both of these challenges. Our specific goals and associated approaches are: (1) To improve statistical methods for estimating transmission and contact networks from diverse biological data, based on a novel integration of dual infection and selection graph (DISG) and exponential random graph model (ERGM) approaches. (2) To generalize from a particular realization of a transmission network process to a probabilistic model for the underlying contact-generating process. This phase will be based on refinements of ERGMs to allow for incomplete network data, as estimated in Goal 1. (3) To validate, refine, and generalize the models developed in Goals 1 and 2 using recursive simulation methods. (4) To ground-truth the methods using uniquely detailed data on a host/pathogen system involving a fast-evolving virus (feline immunodeficiency virus, FIV) in wild cougars. Our research will contribute importantly to human health through novel understanding of disease transmission in populations with unique contact and movement patterns, thereby informing public health officials of epidemic risks and potential intervention strategies for both well-characterized and novel infections.
| Carnegie, Nicole Bohme; Krivitsky, Pavel N; Hunter, David R et al. (2015) An approximation method for improving dynamic network model fitting. J Comput Graph Stat 24:502-519 |
| Schweinberger, Michael; Petrescu-Prahova, Miruna; Vu, Duy Quang (2014) Disaster Response on September 11, 2001 Through the Lens of Statistical Network Analysis. Soc Networks 37:42-55 |
| Tibbits, Matthew M; Groendyke, Chris; Haran, Murali et al. (2014) Automated Factor Slice Sampling. J Comput Graph Stat 23:543-563 |
| Hunter, David R; Goodreau, Steven M; Handcock, Mark S (2013) ergm.userterms: A Template Package for Extending statnet. J Stat Softw 52:i02 |
| Vu, Duy Q; Hunter, David R; Schweinberger, Michael (2013) MODEL-BASED CLUSTERING OF LARGE NETWORKS. Ann Appl Stat 7:1010-1039 |
| Stack, J Conrad; Bansal, Shweta; Anil Kumar, V S et al. (2013) Inferring population-level contact heterogeneity from common epidemic data. J R Soc Interface 10:20120578 |
| Hunter, David R; Krivitsky, Pavel N; Schweinberger, Michael (2012) Computational Statistical Methods for Social Network Models. J Comput Graph Stat 21:856-882 |
| Groendyke, Chris; Welch, David; Hunter, David R (2012) A network-based analysis of the 1861 Hagelloch measles data. Biometrics 68:755-65 |
| Schweinberger, Michael (2011) Instability, Sensitivity, and Degeneracy of Discrete Exponential Families. J Am Stat Assoc 106:1361-1370 |
| Welch, David (2011) Is network clustering detectable in transmission trees? Viruses 3:659-76 |
Showing the most recent 10 out of 21 publications
