Workshop on Computational Toxicology

Workshop on Computational Toxicology

The goal of this workshop is to discuss how in vitro and in silico models of tissues can advance our understanding of toxicology. The workshop will start with talks from scientists from the Environmental Protection Agency, the Hamner Institute, Indiana University and Virginia Tech. The rest of the workshop will comprise of multiple sessions on using CompuCell 3D to build computational models of engineered tissues and their responses to environmental chemicals.

Workshop Schedule
September 5, 2013
Kelly Hall 310
Time Title Speaker Institute
8:15am-8:30am Breakfast
8:30am-8:45am Welcome remarks Roop Mahajan Director, Institute for Critical Technology and Applied Science
Virginia Tech
8:50am-9:20am Virtual Liver EPA STAR Consortium Imran Shah National Center for Computational Toxicology
Environmental Protection Agency
9:25am-9:55am Multi-Scale, Virtual Tissue Simulations of Developmental Toxicity, Development, and Developmental Diseases James Glazier Director, Biocomplexity Institute
Indiana University
10am-10:30am Unraveling the wiring of toxicity pathways using computational networks Patrick McMullen The Hamner Institute
10:30am-10:45am Coffee break
10:45am-11:15am High Performance Computing Based Networked Immunology Stephen Eubank and Madhav Marathe Deputy Directors
Network Dynamics and Simulation Science Laboratory
Virginia Bioinformatics Institute
Virginia Tech
11:20am-11:50am Pathway Analysis of ToxCast Data Allison Tegge Department of Computer Science
Virginia Tech
12pm-1:30pm Lunch
1:30pm-2:00pm A Liver Centric Multiscale Modeling Framework of Xenobiotics Jim Sluka Biocomplexity Institute
Indiana University
2:05pm-2:35pm CompuCell3D and Cellular Potts Modelling Julio Belmonte Biocomplexity Institute
Indiana University
2:35pm-5:00pm Modeling session with CompuCell
September 6, 2013
Kelly Hall Cafe X
8:30pm-10:30pm Modeling session with CompuCell
10:30am-10:45am Coffee break
10:45am-12:30pm Modeling session with CompuCell
12:30pm-1:30pm Working lunch
1:30pm-3:00pm Modeling session with CompuCell
3:00pm-4:00pm Discussions

Abstracts

Multi-Scale, Virtual Tissue Simulations of Developmental Toxicity, Development, and Developmental Diseases

James A. Glazier
Biocomplexity Institute and Department of Physics
Indiana University

Our ability to integrate molecular and genetic information to make biomedically meaningful predictions at the organ or organism level is still limited, because of the difficulty of predicting the emergent properties of tissues from cells' molecular signatures. For the past 15 years, we have been developing computational tools and approaches to bridge the gap between molecule and physiological outcome. Our opensource CompuCell3D modeling environment enables rapid specification and refinement of complex biomedical computer simulations that combine subcellular molecular reaction-kinetics models, the physical and mechanical behaviors of cells and the longer range effects of the extracellular environment. I will illustrate these approaches in computer simulations of: 1) the effects of environmental toxins on intersegmental angiogenesis in zebrafish (a potential medium-to-high throughput screen for human developmental toxicity) and 2) polycystic kidney disease. I will specifically focus on the types of questions that simulations can address and the types of experimental data required for their development and validation.

High Performance Computing Based Networked Immunology

Stephen Eubank and Madhav Marathe
Network Dynamics and Simulation Science Laboratory
Virginia BioInformatics Institute
Dept. of Public Health and Computer Science
Virginia Tech

We will discuss a novel approach based on a combination of high performance computing and network science to understand the cellular interactions in an immune system. We will focus here on gut immunity and discuss ENteric Immunity Simulator (ENISI). The work is being carried out under the NIH funded MIEP project (http://www.modelingimmunity.org/). ENISI is a modeling system for the inflammatory and regulatory immune pathways triggered by microbe-immune cell interactions in the gut. ENISI allows immunologists and infectious disease experts who can test and generate hypotheses for enteric disease pathology and identify potential treatment strategies that reduce inflammation induced damage through experimental infection of an in silico gut.

Joint work with Keith Bisset, Md. Maksudul Alam, Josep Bassaganya-Riera, Adria Carbo, Xinwei Deng, S. Raquel Hontecillas, Stefan Hoops, Yongguo Mei, Katherine Wendelsdorf, and Jae-Seung Yeom.

Pathway Analysis of ToxCast Data

Allison Tegge
Department of Computer Science
Virginia Tech

Our population is exposed to toxicants on a daily basis. Previous studies have focused on identifying toxic exposure levels, but determining these levels is non-trivial and remains an outstanding challenge. Furthermore, determining the toxic exposure level does not provide the mechanism of toxicity. Recently, the EPA has started the ToxCast initiative that uses high-throughput chemical response data to try to understand the mechanisms of response from the cell. Currently, ToxCast includes high-throughput chemical response data of over 225 human and rat genes after exposure to 59 toxicants.

To better understand the mechanisms of response after exposure, we sought to identify probable signaling pathways that connect the responsive receptors to transcription factors that represent the cellular signaling mechanisms perturbed by each chemical. Using Linker, a computational approach for reconstructing signaling pathways developed in our group, a background protein-protein interaction network, the responsive receptors, and the responsive transcription factors, we computed the cellular response network as a result of chemical exposure. For each chemical, we re-weighted the interaction network such that proteins in the neighborhood of the responsive receptors had a higher weight, and those proteins more distant have lower weights. Then, we identified multiple highest scoring paths connecting the responsive receptors to transcription factors. These paths are likely to be representative for how a chemical may perturb normal cellular signal transduction. The proteins in the top 250 predicted signaling response paths after exposure to Bisphenol-A belong to several pathways that have been previously been implicated in the literature. We have also made predictions on less well-studied chemicals included in ToxCast. The literature confirmed predictions from the Bisphenol-A analysis provide confidence to our predictions for these less understood chemicals.

A Liver Centric Multiscale Modeling Framework of Xenobiotics

Jim Sluka
Biocomplexity Institute
Indiana University

Pharmacological and toxicological processes occur across a wide range of scales and include multiple organ systems. A true Systems Biology pharmacological model must include submodels that cover multiple scales and multiple tissues relevant to human medicine and toxicology. We describe a multiscale modeling framework for xenobiotic pharmacology, metabolism and toxicity that incorporates a PBPK whole body representations, tissue level (multicell) behaviors and subcellular signaling and metabolic pathways. At the PBPK level we represent uptake, distribution and excretion of a xenobiotic. At the multicell level we model tissue level behaviors such as cell proliferation and death. At the subcellular level we represent signaling, metabolic and gene expression processes. The models at each of these three scales communicate with the model(s) at higher or lower scales.

We link existing open source tools into an aggregate model and avoided building a single monolithic tool. This approach allows us to leverage not only preexisting tools but also preexisting models at the individual scales. We implement PBPK whole body models as Systems Biology Markup Language (SBML) models, multicell and tissue scale as CompuCell3D (CC3D) models and subcellular signaling, metabolism and gene expression models as SBML. Copies of the subcellular SBML models are embedded into each of the cells in the CC3D model. In turn, the CC3D model is a compartment (tissue) in the whole body PBPK model. CC3D uses the SBML ODE Solver Library (SOSlib) for solving systems of ODEs. CC3D controls the models at the higher and lower scales since both the whole body and subcellular models are expressed in SBML, which can be controlled by CC3D. The individual models at each of the three scales can generally be run as a standalone (single scale) model. This facilitates validation at the individual scales and makes it easy to incorporate preexisting single scale models developed by others.

As an example of linking these existing tools and models, from the whole body to the subcellular, we present a model of Acetaminophen ADME, pharmacological action and liver toxicity.

CompuCell3D and Cellular Potts Modelling

Julio Belmonte
Biocomplexity Institute
Indiana University

The Cellular Potts model, also known as the Glazier-Graner-Hogeweg (GGH) model, is one of the most used mathematical models to describe and simulate cell and tissues. The model was originally developed to simulate just a few cellular behaviors such as growth and adhesion, but over the years it has been extended to simulate a multitude of cell behaviors, including diffusion fields, subcellular networks and metabolic pathways, making it an ideal tool for multi-cell, multi-scale modelling. In this talk I will present the basics principles of the Cellular Potts model and show a collection of models that illustrates all its capabilities. Finally I will formally present the CompuCell3D software, a user-friendly, open-source simualtion enviroment that allows for fast development of these kind of models.