Projects
From epicenter_wiki
Theory and Optimization of Robust Complex Networks
PI: Caterina Scoglio, co-PI: Rob Kooij
The proposed research focus on proving that network robustness is intrinsically related to network topology. The objective of the proposed research is to develop (1) analytical tools to select proper statistical graph-theoretical metrics to describe the level of connectivity of the network and consequently its robustness with respect to massive failure and spread of malware, and (2) practical engineering tools to design and manage robust and resilient complex networks.
Optimizing antimicrobial treatment regimens to minimize the spread of resistance.
PI: Ronette Gehring
Antimicrobial drugs are powerful instruments for the treatment and management of bacterial disease. The inappropriate use of these drugs can, however, increase the prevalence of resistance in bacterial populations. Not only does the use of antimicrobial drugs selectively promote the survival of resistant individuals, but it also increases the horizontal spread of genetic elements encoding for resistance within the population. This leads to infections that fail to respond to treatment, resulting in death, prolonged illness and longer hospital stays estimated to cost the US healthcare industry up to $7 billion annually. Our proposal seeks to develop a mathematical model that links the dynamic concentrations of an antimicrobial drug in the body with changes in the size and composition of target bacterial populations over time. The overlay of pharmacokinetic, epidemiological and contact network models make it possible to incorporate multiple factors, including the rate at the antimicrobial drug kills susceptible bacteria, the relationship between drug concentrations and the rate of transfer of genetic resistance elements between individuals, as well as characteristics of the site of infection that may affect opportunities for contact between donor and recipient bacteria. Model parameters will be calculated independently from in vitro experiments. Predictions will be compared with results from in vivo models of clinical infection. The aim of this project is to develop this model that it may be used as a tool to optimize treatment regimens.
Predicting rabies exposure risk and evaluating intervention methods with spatially explicit, dynamic contact networks
PI: Sam Wisely
The specific aim of this proposal is to develop a spatiotemporally explicit, individual-based model of landscape epizootiology, using dynamic contact network modeling developed at EpiCenter. The investigators will use this model to test hypotheses of how ecological patterns of urbanization, including resource distribution, habitat degradation, and physical barriers, shape the epizootiological processes of rabies with a primary focus towards the prediction of public health risk and the effect of disease intervention methods. To develop and parameterize this model, the investigators will compare and contrast the endogenous and exogenous mechanisms of transmission in two distinct strains of rabies that occur in a common host species, the striped skunk. The intellectual merit of the proposal lies in the construction of a spatio-temporal model of epizootiology. The simulation software developed during this project will allow investigators to forecast changes in disease risk associated with the changing rural-urban interface, and explore different disease management scenarios in order to develop cost efficient and effective disease control measures. The strength of the model relies on a hierarchical approach to understanding disease ecology. At the organismal level, the investigators will compare viral transmission properties and characterize the molecular attributes of two virus variants. At the population level, they will characterize seasonal changes in patterns of habitat affinities in striped skunk to estimate contact rates and dispersal probabilities along a clinal ecotone of urban and rural ecosystems. At the regional level, they will use population and landscape genetic approaches to estimate migration rates and identify habitat corridors through which both host and pathogen may flow.
A novel diffusion model describing the absorption of chemicals through complex biological membranes
PI: Deon van der Merwe
Our skin is the point of contact between our bodies and a multitude of xenobiotic chemicals in our environment. The large surface area and accessibility of the skin not only makes it vulnerable to toxic chemicals, but also makes it a convenient site for the application of pharmaceuticals and cosmetics. The effects that these chemicals may have, whether it be toxic effects or beneficial effects, depend on the degree to which they are absorbed into the body. The skin is a highly complex biological membrane that evolved to protect us from excessive water loss and fluctuations in external temperature. It also serves as a barrier to mechanical and chemical insults. Most of the barrier properties of the skin depend on the outer skin layer, called the stratum corneum. The stratum corneum consists of layers of keratinized cells embedded in a semi-crystalline, layered lipid matrix. The lipid matrix is the most important route of absorption through the skin for most chemicals. Current models view the lipid matrix as a homogenous medium through which chemicals diffuse passively from high to low concentration according Fick’s first law of diffusion. This model predicts that the fraction of a chemical that is absorbed through the skin over a specified period of time is constant, and will remain constant at various concentrations of chemical applied to the skin surface. It also predicts that chemicals will spread through the lipid matrix in a smooth continuum from high to low concentration. Both of these predictions are contradicted by experimental evidence. The fraction of the applied dose of a chemical that is absorbed per unit of time diminishes as the dose concentration is increased; and chemicals tend to penetrate the lipid matrix in an irregular pattern, concentrating in some areas of the lipid matrix, while being present at much lower concentrations in other areas. We propose an alternative model for describing chemical movement in the stratum corneum lipid matrix based on successive partitioning through multiple compartments that describe the physical-chemical characteristics of the lipid matrix. This model offers the opportunity to describe the absorption of chemicals through skin in a manner that closely resembles actual diffusion processes and experimental results.
Network models for soybean rust epidemics: Adapting to aerially-dispersed pathogens
PI: Karen Garrett
Modeling plant disease epidemics at large scales calls for several adaptations of network models. For example, information about soybean rust is typically available from sentinel plots that function to represent a larger area such as a county. Furthermore, many plant pathogens are capable of long-distance aerial dispersal, so that distant nodes may be connected with a small but non-zero probability.
Complex network approach to epidemic spreading in rural regions
PI: Todd Easton, co-PI: Walter Schumm
The overarching goal of this research is to develop optimized guidelines that administrators can use to establish procedures and realign resources to help mitigate the effects of the spread of infectious diseases in rural regions such as Western Kansas, caused by natural causes or malicious attacks. This project studies the peculiar contact network of rural regions, and develops a simulation tool with multiple compartments running on the contact network.
Individual-based model for simulation of influenza epidemic spread in large cities
PI: Valeriy Perminov
It is well known that transmission of the influenza infection takes place mainly through close contacts (in kindergartens, schools, households, and workplaces) and to a lesser extent through casual contacts (in cafes, cinemas, public transport, etc.). The traditional mean field approach to epidemic spread studying based on a wellmixed assumption in the whole population or in age groups can not take into consideration these peculiarities. On the contrary individual-based models (IBMs) may take into account these (and many, many other) details. In such models population is an aggregate of individuals. Each individual is characterized by a set of properties. These properties define an individual’s social and medical features and some part of them can change with time.
