Guest Speaker
Using Agent - based Modeling to Study Complex Systems
Virginia A. Folcik Nivar
Dept. of Internal Medicine; Div. of Pulmonary, All
The Ohio State University Medical Center
Dec 3 2009 3:30PM
480 Dreese Labs
All interested parties are welcome to attend.
Refreshments will be available in the presentation
Abstract:
Complexity science is expected to come of age in the 21st century. It began in the 1990's in various disciplines, including the social sciences, computer science, physics, ecology and entomology. It was initially described as a form of "Artificial Life." Some of the pioneers formulated a definition of complex systems by observing ants in their natural habitat for long periods of time. In the last ten years, cell biologists and immunologists have adopted these ideas and have applied them to cellular systems.
Complex systems are composed of interacting units with rules for behavior that depend on what they can sense in their immediate environment. The types of things that these units or agents can sense are other agents or signals in the environment. The signals can be substances emitted by other agents, or changes made to the environment that an agent can recognize as a pattern. Agent-based modeling is a computational method that was invented to model complex systems. It is a bottom-up method, with the information about the system gathered from the perspective of the agents that make up the system. Behavior is programmed into the agents, represented as objects with variables and methods. The agents are put into an environment (another object, a container) where they can interact. The agents have behavioral states, depending on what they experience over time. The combination of the agents and their behavior in the environment makes up the model of the system. Properties of the system as a whole emerge from the bebehavior of the agents.
In one of my agent-based models, the Basic Immune Simulator, the agents are the cells of the immune system, the signals are cytokines and chemokines (molecules emitted by cells), and there are three environments. The environments represent: 1) a non-specific tissue, 2) the lymph nodes/spleen, and 3) the blood. These are the three environments where the cells of the immune system interact during an immune response. In another version, the tissue environment represents the lung, and this model is being created to study lung disease. In a completely different model the agents are healthcare workers and the environment is the Medical Intensive Care Unit. This model is being created with the collaboration of my colleagues in an Innovation Group.
The science of complexity will flourish when more computer scientists learn about agent-based modeling, and more scientists in other fields learn about complexity and the basic logic of computer science. The future of research is cross-disciplinary interaction among scientists, which indeed needs to be instilled in the minds of future scientists. I am using agent-based modeling to work on solving problems in medicine that have defied traditional approaches to research. I am involving undergraduate students, all potential future scientists, so that they can benefit from trans-disciplinary research experience. These are the main goals of my work.
Bio: Dr. Folcik earned her bachelor's degree in Biology and Medical Technology from Cleveland State University (CSU) in 1986. Then, working in the laboratory of Martha K. Cathcart, Ph.D. in the department of Cell Biology of the Lerner Research Institute of the Cleveland Clinic Foundation she continued on to earn her Ph.D. (from CSU) in Regulatory Biology. Her thesis work involved studying the role of the immune system in low-density lipoprotein (LDL) oxidation and atherosclerosis, and the modulation of monocyte-mediated LDL oxidation by cytokines. Dr. Folcik was a post-doctoral fellow for four more years studying the contribution of lipoxygenases to lipid oxidation in atherosclerosis. This work was supported by fellowship grants awarded to Dr. Folcik by the American Heart Association and the National Institutes of Health (NIH), National Research Service Award program. Dr. Folcik then went on to study the autoimmune disease Multiple Sclerosis in a murine model called Experimental Autoimmune Encephalomyelitis, with Dr. Richard Ransohoff in the department of Neurosciences Research of the Lerner Research Institute. She analyzed the effects of phosphodiesterase inhibitors on the blood-brain barrier and disease outcome in the mice. After completing these studies, Dr. Folcik pursued a degree in Computer Science and Engineering, earning a bachelor's degree at OSU in 2005. In 2002, Dr. Folcik began working with Dr. Charles G. Orosz in Transplant Surgery to create an agent-based model of the immune system called the Basic Immune Simulator. After Dr. Orosz passed away in August, 2005, Dr. Folcik continued to work on the model. She joined the staff of the Pulmonary, Allergy, Critical Care and Sleep Medicine division in the department of Internal Medicine with the support of Dr. Clay Marsh and continued this work. She currently has R21 funding (as of 6/2009) from the NIH as the principal investigator of a project entitled "Agent-based modeling to reveal mechanisms of idiopathic interstitial lung disease." This project involves extending the Basic Immune Simulator to study lung disease, and employs five undergraduate computer science students.
http://digitalunion.osu.edu/r2/summer06/sass/
http://www.complex-systems.wikidot.com
Host: Bruce Weide
