Filtering, segmenting, and modeling structures in tomograms is often out of necessity done by hand and eye with interactive software. Where algorithmic methods are feasible, interactive analysis is used to determine input parameters and assess computational results. Achieving optimal results requires facile interactive data exploration. I'll demonstrate how software we develop for visualizing tomograms can be used to explore possibilities for subtomogram averaging in specific data sets, showing recent improvements to our UCSF Chimera visualization package in the areas of filtering, slicing, masking, boxing particles, symmetry evaluation, missing-wedge compensation and fly-through animation.

3D animation software is a powerful tool that allows users to synthesize data from diverse sources to create dynamic models of molecular and cellular mechanisms. Molecular animations are useful not only as a means to better communicate scientific findings and hypotheses, but also give researchers insight into the processes they study. I will show and discuss several examples of visualizations constructed using Autodesk Maya and UCSF Chimera, including animations of clathrin-mediated endocytosis and protein translocation through the ER.

UCSF Chimera is a program for interactive molecular graphics and modeling. It provides standard graphics features as well as more unique, domain-specific tools; the menu and command-line interfaces provide a rich and overlapping set of functionality. The Introduction to Chimera shows frequently used coloring and display options, including molecular representations such as ribbons, "pipes and planks," surfaces, and abstract renderings of nucleotides. Other general features shown are distance measurements, bond angle rotations, H-bond identification, and display of the corresponding amino acid and/or nucleotide sequences. Attributes such as B-factors and hydrophobicities can be rendered visually with colors, atomic radii, and "worm" thickness. Chimera includes detailed user documentation and is available for Windows, Linux, Mac OS X (with X11), IRIX, and Tru64 Unix. Chimera is free for academic, government, and non-profit use and can be downloaded from http://www.cgl.ucsf.edu/chimera.

Biological models are becoming more and more complex. Models of biochemical kinetics accounting for dozens of different species are a norm; models accounting for hundreds of species and reactions are no longer rare. Whether a modeller builds a new model or modifies an existing one, manually specifying a list of species and reactions is error-prone and slow. Though there are a variety of software tools for simulation of biochemical kinetics, very few tools provide modelling capabilities beyond manual specification of each and every model element. Here we demonstrate BioNetGen software that provides a mechanism to specify a model in the form of bio-molecular interaction rules that automatically generate a biochemical reaction network. We will demo several versions of BioNetGen, including downloadable software (http://bionetgen.org ) and web-application BioNetGen@VCell ( http://vcell.org/bionetgen ) that is developed as a part of the Virtual Cell ( http://vcell.org) modeling and simulation environment.

How can we combine electrophysiology and biochemistry in a virtual Purkinje neuron? This problem typifies how neurobiology can benefit from interweaving computational neuroscience and electrophysiology with biochemistry and systems biology. Membrane potential can successfully be modeled in the Virtual Cell, but the complex geometry of the cell poses some computational limitations. Building on our previously published biochemical models of calcium dynamics, we created electrophysiological models in Virtual Cell, based on a model (Miyasho et al, 2001, Brain Research 891:106-15) available in NEURON, a popular neuroscience simulation software. Because the software enforces an equipotential within individual compartments, we linked multiple compartments together, in a Virtual Cell MathModel, defining the electrical coupling between them with expressions of current density as a function of resistivity and compartmental geometrical parameters. We created similar simulations in NEURON for comparison. In addition, we modified and applied an accepted reduction method to branched neuron test geometries to obtain models with fewer compartments so as to produce tractable Virtual Cell models. In these models, most of the dendritic tree was subject to reduction, but we retained the neurons' electrical integrity along a specified path from spine to soma. We conserved axial resistivity and created equivalent reduced cylinders with membrane resistance and capacitance adjusted for the reduction in surface area. We found that an identical current injection at the soma only in (i) an isolated soma in a Virtual Cell BioModel, and (ii) multiple connected compartments in a Virtual Cell MathModel was able to reproduce a train of action potentials and membrane potential propagation, as in NEURON. Further, similar results were obtained for single action potentials and trains of action potentials due to current injections at the soma in the full and reduced versions of a small test model with and without an attached preserved path. Finally, an alpha function applied at the spine to represent synaptic stimulation gave similar results in the spine and the soma in both the full and reduced small test models. Thus, Virtual Cell faithfully modeled propagation of membrane potential between compartments, and our reduction modification reproduced electrical properties between a preserved dendrite branchlet and the soma. (Supported by NIH P41 RR013186 and U54 RR022232)

The NIH Resource for Macromolecular Modeling and Bioinformatics is located at the Beckman Institute for Advanced Science and Technology, on the University of Illinois at Urbana-Champaign campus. Advanced computational methods developed at the Resource, realized in the software packages NAMD and VMD, offer today new microscopic views of cellular structures and processes by combining multimodal experimental data (e.g., from light, electron, and atomic force microscopy as well as X-ray scattering and NMR structure analysis) with physical modeling (e.g., quantum chemistry and multi-scale molecular dynamics simulations). Structures and processes can be visualized over wide spatial and time scales. Applications include integral views of proteins exerting forces on cells or sculpting cellular membranes, of ribosomes in different functional states, and of entire photosynthetic units absorbing sun light and charging the cellular membrane. This "computational microscope", besides offering unprecedented views, adds static and dynamic information not available from experiment alone, e.g., in case of disordered systems like membranes, or adds dynamic information where experiment offers only static images.

The software packages NAMD and VMD, which comprise the computational microscope, are both distributed free of charge with source code. These software tools facilitate the discovery process from analysis, through modeling, to visualization of the molecular apparatus in biological cells:

• NAMD, recipient of a 2002 Gordon Bell Award, is a parallel molecular dynamics code used regularly to simulate systems of 1,000,000 atoms and beyond on both large supercomputers and inexpensive Linux clusters.
• VMD is a molecular visualization program for displaying, animating, and analyzing large biomolecular systems using hardware-accelerated 3-D graphics and built-in scripting.

A tight coupling between cell structures, ionic fluxes and intracellular Ca2+ transients underlies the regulation of cardiac cell function. We have used multi-scale meshing and numerical approaches and tools (GAMer, FETK, MCell) to investigate these complex interactions at nanometer (dyadic cleft) and micrometer (single t-tubule) scales in rat ventricular myocytes. The realistic model geometries have been extracted from both light and electron microscopy images. Preliminary results suggest that cardiac cell function is tightly regulated by the localization of Ca2+-handling proteins and strongly relies on the presence of mobile and stationary Ca2+ buffers and cell geometry.

With the advent of Cloud computing, the concepts of Software as a Service (SaaS), where vendors provide key software products as services over the internet that can be accessed by users to perform complex tasks, and Service as Software (SaS), where customizable and repeatable services are packaged as software products that dynamically meet the demands of individual users, have become increasingly popular. Both SaaS and SaS models are highly applicable to scientific software and users alike. Opal2 is a toolkit for wrapping scientific applications as Web services on Grid and cloud computing resources, which provides a mechanism for scientific application developers to expose the functionality of their codes via simple Web services APIs, abstracting out the details of the back-end infrastructure. Complex services may be combined via customized workflows for specific research areas and enables the packaging of complete work environment to be distributed as virtual machine images. The new plug-in architecture for Opal2 will be demonstrated, with available plug-ins for CSF4, Globus-GRAM, DRMAA and fork.