![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
2007 CompTox Forum
Abstract - The Virtual Liver Project at the U.S. EPA's National Center for Computational Toxicology and Its Implications for the EPS Mission to Protect Human Health
Imran Shah, Ph.D.
Computational Systems Biologist
U.S. EPA, Office of Research and Development
National Center for Computational Toxicology
Mail Code: B205-01, Research Triangle Park, NC 27711
Phone: 919-541-1391
E-mail: shah.imran@epa.gov
Computational modeling of biological systems has, to date, been exemplified by physiologically-based pharmacokinetic (PBPK) models, clonal growth models of cancer, and, to a lesser extent, models of cellular signaling and gene-regulatory networks. These models tend to focus on specific levels of biological organization and contain relatively little information about structure and process across levels of organization. The "omics" revolution in the biology laboratory and parallel developments in computer hardware and software technologies are now enabling much more ambitious modeling of biological systems, including the development of virtual tissues where the relevant biology is described across multiple levels of organization. The virtual liver project (VLP) at the U.S. Environmental Protection Agency's (U.S. EPA's) National Center for Computational Toxicology is integrating existing knowledge about hepatic biology and mechanisms of toxic action, with the long-term goal of providing a capability for predicting dose-response and time-course behaviors for biochemical and structural changes in response to toxicant exposures. In the shorter term, the project is focusing on collection of data on hepatic biochemistry, physiology, and architecture, and on the development of a database and query tools. Also underway is an initial effort to describe the regulation of induction of xenobiotic metabolizing enzymes that are often implicated in toxicological mechanisms. Construction of the VL can be compared to working on a very large jigsaw puzzle. Initially, we have mostly unorganized information and data gaps (i.e., missing pieces of the puzzle). With time, existing information will be integrated (the existing pieces will be fit together to the extent possible) and the developing model will help to direct new research to fill the data gaps (create missing pieces). This process will continue iteratively. The developing VL model will thus traverse multiple cycles of development and model-directed data collection. With time, we expect the VL to contribute significantly to our understanding of how perturbation of normal biological processes leads to toxicity and to provide quantitative predictions of dose-response in support of risk assessment.