HARMONY 2019 poster abstracts

This list is not in any particular order. Last updated 2019-03-20.

Model Engineering - Building Biomedical Models Using Principles from Software Engineering

Joseph L. Hellerstein (eScience Institute, University of Washington)
Woosub Shin (eScience Institute, University of Washington)
Herbert M. Sauro (Bioengineering Dept., University of Washington)

Biomedical models face issues with reproducibility, readability, and reuse. Software grew to be a multi-trillion dollar industry by addressing these issues. Model Engineering is about adapting tools and techniques from software engineering to improve biomedical models. The poster describes our current efforts in pursuing the Model Engineering research agenda.

Website with more information: https://f1000research.com/articles/8-261/v1

Simulate Spatially Resolved Rule-based Models with Simmune

Fengkai Zhang (Laboratory of Immune System Biology, NIAID/NIH, USA)
Martin Meier-Schellersheim (Laboratory of Immune System Biology, NIAID/NIH, USA)

Biochemical modeling efforts increasingly take advantage of rule-based modelling approaches to construct biochemical reaction networks based on the specification of interactions of individual molecular components. In addition to biochemical interactions, cellular signaling processes depend on the spatiotemporal distributions of their molecular complexes. We developed the Simmune software suite to provide a platform for simulating spatially resolved rule-based models. The Simmune Modeler provides a flexible visual interface to allow users to specify models iconographically through the definition of molecular components and their interactions and of the reaction-induced modifications of molecular domains. The same iconographic language is used by the Simmune Network Viewer to render network representations of the models while the Simmune Cell Designer permits defining cellular morphologies, including sub-cellular structures, such as nuclei. For spatially resolved simulations of cellular biochemistry Simmune can handle even dynamic morphologies whose changes are driven by the biochemistry. Moreover, Simmune offers parameter scanning and fitting procedures and permits exporting and importing rule-based models in a standardized SBML Multi format. Simmune is available from the NIH website. This work is supported by the intramural program of the NIAID, NIH.

Website with more information: https://www.niaid.nih.gov/research/simmune-project

StochSS: A Next-Generation Toolkit for Simulation-Driven Biological Discovery

Linda Petzold (University of California Santa Barbara)
Brian Drawert (University of North Carolina Asheville)
Andreas Hellander (Uppsala University, Sweden)
Michael Hucka (Caltech)

We describe StochSS, a novel and easy to use Software-as-a-Service offering for quantitative modeling of biochemical networks capable of seamless deployment in public or private cloud environments, as well as your laptop. StochSS integrates advanced algorithms for modeling biochemical systems on multiple levels, ranging from simple deterministic ordinary differential equations (ODEs), through discrete stochastic models simulated via the Stochastic Simulation Algorithm, all the way to detailed spatial stochastic models capturing stochasticity as well as molecular movement and realistic cellular geometries. In addition to its core simulation capabilities, Next-Gen StochSS will have two new features: the Model Development Toolkit (MDT) for constructing models from time-series data, and the Model Exploration Toolkit (MET) for exploring the parameter space to discover the qualitatively distinct behaviors the model can yield, within the space of uncertain parameters. We welcome your input!

Website with more information: http://stochss.org

Epithelial Modelling Platform: A Tool for Investigating Hypotheses through Discovery and Assembly of Computational Models of Epithelial Transport

Dewan Sarwar (Auckland Bioengineering Institute, University of Auckland)
Koray Atalag (Auckland Bioengineering Institute, University of Auckland)
Peter Hunter (Auckland Bioengineering Institute, University of Auckland)
David Nickerson (Auckland Bioengineering Institute, University of Auckland)

Scientists often leverage computational models of biological systems to investigate hypotheses which are difficult or prohibitively expensive to achieve experimentally. Such investigations are best achieved by utilizing suitable computational models, reusing existing validated models where possible and creating novel models consistently as needed. This requires tools which enable the discovery and exploration of existing models matched with assistance in constructing and testing new models. Enabling scientists or clinicians to use such a tool by describing their requirements in a manner familiar to themselves greatly improves the accessibility of the tool.

We have developed a web-based tool, the Epithelial Modelling Platform, for scientists and clinicians to discover relevant models and then assemble these into a novel model customized for investigating their hypotheses. While our tool specifically focuses on epithelial transport, by utilizing relevant community standards and publicly accessible knowledge repositories, it is extensible to other areas of application. The platform abstracts underlying mathematics of the computational models and provides a visual environment which mimics biological phenomena of an epithelial cell.

Beyond the mathematical models, we have implemented a feature to discover existing simulation experiments which match the features of the novel models users create. By executing these simulation experiments with the novel models and comparing to previous model predictions and/or experimental or clinical observations we are able to provide the user with some measure of verification that their model matches, or doesn’t match, existing knowledge captured in the various repositories utilized.

Website with more information: https://github.com/dewancse/epithelial-modelling-platform

netplotlib: a simple package for visual analysis of reaction network models

Kiri Choi (Department of Bioengineering, University of Washington)
Herbert M. Sauro (Department of Bioengineering, University of Washington)

netplotlib is a Python-based package which works as a simple extension to matplotlib and NetworkX to analyze and generate specialized reaction network diagrams with ease. netplotlib supports models written in SBML or Antimony. The package provides network analysis capability while being highly customizable. We plan to integrate our package as an extension to libsbml-draw visualization library which supports SBML layout and render specifications.