Competitive Renewal of Development, Improvement and Extension of the Tissue Simu

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): During two-and-a-half years of funding under NIH GM077138, the EPA's ComTox program, the Dundee computational oncology effort and an increasing number of developmental biologists and tissue engineers have adopted CompuCell3D (CC3D) as a modeling platform. CC3D has become the most widely accepted standard for multi-scale, multi-cell (MSMC) simulations of developmental phenomena and related diseases, with more than 100 trained users, because its open-source combination of the multi-cell capabilities of the Glazier-Graner-Hogeweg (GGH) model, the SBML-compatible subcellular biochemical networks, and continuum and finite-element modeling of tissue-level phenomena, sophisticated user interfaces and scripting capabilities allow rapid construction of useful biomedical models with tunable levels of modeling detail. Motivated by the requests of our increasing number of translational users, this competing renewal will develop a parallel (Graphical Processing Unit (GPU), multi-core and cluster) CC3D2 with fluid-flow support providing up to a hundred-fold increase in speed over CC3D plus the ability to run very large-scale simulations (many cm3, 107-109 cell, whole embryo/organ) with a simple migration path from laptop to diverse parallel architectures (Specific Aim 1). It is significant, because CC3D2 will transform the power of MSMC modeling to allow the development of the detailed, verifiable models required for future clinical-research applications, achieving the goal identified by the NIH-led Interagency Modeling and Analysis Group (IMAG) group: "the development of open source, multi-scale biological simulation environments which run both on single processors and parallel computers and which ..., [permit] users to select the level of simulation detail without further modifying their simulations." It is innovative because it combines the existing expertise of the CC3D development team and the extensive GPU and parallelization experience of the D'Souza group to provide a robust parallel MSMC environment and the first GPU-based implementation of the GGH methodology to allow simulations which were formerly unachievable. As requested by our clinical and biomedical users, CC3D2 will provide an innovative, fast and biologically- intuitive approach to model design, with a novel Cell Behavior Model Specification Language (CBMSL) (Specific Aim 3) and the first sharable graphical model definition available for MSMC (Specific Aim 2). CBMSL will allow researchers to focus on biology rather than computational details and greatly facilitate model cross-validation and sharing, providing an important use-case for future MSMC model-description standardization efforts. Graphical workflow control and enhanced user support and documentation (Specific Aim 4) will further improve CC3D2 usability and increase user adoption.