Seg3D is a free, open-source volume editing and processing tool under development by the Center for Integrative Biomedical Computing at the University of Utah Scientific Computing and Imaging (SCI) Institute. Seg3D combines a flexible manual segmentation interface with powerful higher-dimensional image processing and segmentation algorithms from the Insight Toolkit, such as nonlinear image denoising, edge-detection, statistical flood filling, morphological operations, and level-set segmentation algorithms. Users can explore and label image volumes using configurable, orthogonal slice view windows, and (coming soon) 3D volume rendering. Segmentations are displayed as isosurfaces in a 3D scene as they are created. An important feature of Seg3D is that the outputs of automatic segmentation algorithms are fully editable in the manual interface, and may be subsequently processed by any number of additional filtering operations. Label masking logic allows users to map ontological relationships between objects in a scene to the segmentation process to enforce concepts such as "containment" and "adjacency".

Seg3D is a general-purpose tool. It is currently undergoing testing for use in segmenting scenes in electron microscope tomography volumes by the UCSD National Center for Microscopy and Imaging Research. Seg3D is also being tested on heart image data at the SCI Institute to create models for bioelectric field research. At SCI, we are also producing bone segmentations of MicroCT scans of mice from the Mario Capecchi Laboratory (University of Utah). Other research groups in the biomedical imaging community are also evaluating Seg3D, including the National Alliance for Medical Image Computing and NIH Heart Lung and Blood.

We will use the UCSF Chimera program ( www.cgl.ucsf.edu/chimera) to look at human RNA polymerase II (Kostek 2006), a 12 protein complex that synthesizes RNA from a DNA template. This live demonstration will examine electron microscopy 3-dimensional maps of two conformations and a yeast atomic model showing how to morph one map to another, color principle features of the maps, fit the atomic model into a map, modify the atomic model to better fit the map, morph between atomic models, and explore protein sequence differences between human and yeast. Questions that stump the presenter are welcome!

UCSF Chimera is an interactive molecular graphics program for analysis of proteins, nucleic acids, volumetric and sequence data. and for creating publication images. The density map display and analysis capabilities are being developed for studying single particle reconstructions and EM tomography. Chimera runs on Windows, Mac, and Linux operating systems, is free for academic use, and is developed by the Resource for Biocomputing, Visualization and Informatics.

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.

UCSF structureViz is a Cytoscape plugin that links the visualization of biological networks (and biological relationships expressed as networks) provided by Cytoscape with the visualization and analysis of macromolecular structures and sequences provided by UCSF Chimera. structureViz provides commands to open structures in Chimera, manipulate those structures, and align open structures using Chimera's Sequence/Structure tools. n order to load a structure associated with a node, the Protein Databank (PDB) identifier (or identifiers if there are more than one) must be present as an attribute of that node. Currently, structureViz will look for an attribute named Structure, pdb, or pdbFileName. When a structure is opened, structureViz provides an alternative interface to Chimera: the Cytoscape Molecular Structure Navigator. This interface uses a tree-based paradigm to allow users to select and effect the display of models, chains, and residues, mostly through the use of context menus. Additional commands allow for selection by chemistry (Ligand, Ions, Solvent, Secondary Structure, and in the model context menu, Functional Residues). Users can also take advantage of Chimera's structural alignment capabilities by using the "Align" command. structureViz is available for download at http://www.rbvi.ucsf.edu/Research/cytoscape/structureViz/.

Over the past decade, cryo-electron microscopy (cryoEM) has emerged as a powerful approach to the structural determination of large macromolecular complexes. The National Resource for Automated Molecular Microscopy (NRAMM) was established in 2004 with the mission of developing, testing and applying technology aimed at automating the processes involved in solving macromolecular structures using cryo-electron microscopy. The goal of automation is to facilitate the process of molecular microscopy as well as to expand the scope of accessible problems and push experimental frontiers by making possible investigations that would otherwise be deemed too difficult or high risk because of the considerable effort involved in using manual methods. An additional goal of automation is to enable much higher data throughputs driven by the need to improve resolution for single particle reconstructions by increasing the numbers of particles contributing to the average 3D map. The other major mission of NRAMM is to use the infrastructure developed to open up the sometimes esoteric practices of cryoEM to a much wider group of researchers.

New technologies developed at NRAMM include: a new specimen grid substrate; a robotic EM grid handling system, and Leginon - an automated system for microscope control and image acquisition. The technological developments have been driven by, and in turn enabled, a number of collaborative and service research projects, including for example, reconstructions of a minimal COPII cage, an intact infectious P22 virion, and self-assembling DNA nanoparticles. The automated infrastructure has also made cryoEM more accessible to a wider research community including groups whose primary interests are in chemistry, X-ray crystallography, materials science and the pharmaceutical industry.

Electronic lab notebooks are one of the few types of scientific documents that have not benefited from advances in information technology. Existing scientific databases lack the flexibility and broad applicability required for biophysics. Conversely, electronic lab notebook systems lack the data mining capabilities of true databases. We present EMEN2, an object oriented database designed to overcome many of the limitations of traditional databases. This project is in collaboration with S. J. Ludtke and Hari Damodaran.

One of the bottlenecks in solving high resolution structures of ice-embedded single particles by cryo-electron microscopy is to quickly collect large quantities of high quality data. In collaborating with N. Nakamura at JEOL, Japan, we have built the JEOL Automatic Data Acquisition System (JADAS) for the latest generation of JEOL electron cryomicroscopes. JADAS adopts the concept of a "recipe", or user-defined imaging sequence, which can be customized according to the needs and purpose of an experiment. It can set up parameters of the lens and deflector coils for each operational step in the recipe. This software has been tested for JEM2100 electron cryomicroscope for recording images of ice-embedded single particles.

The type 1 ryanodine receptor (RyR1) is a skeletal muscle Ca2+ release channel that mediates ligand-gated release of Ca2+ from the sarcoplasmic reticulum into the cytoplasm. It is the largest known integral membrane protein, and is composed of four identical subunits with a Mr ~ 565 kDa each. Using single particle cryoEM, we determined the structure of RyR1 in the closed conformation at a 9.6 Å-resolution. Unprecedented structural features in both transmembrane and cytoplasmic regions can be resolved. This project is in collaboration with S. J. Ludtke, M. L. Baker, S. Hamilton, Y. Cong, M. Topf and A. Sali.

Over one hundred actin crystal structures have been reported in the PDB. It is not known how or whether these structures are relevant in the cell. Based on the electron crystallographic structure of acrosomal bundle, we are able to relate some of the actin crystal conformations to each of the 14 actin subunits in the asymmetric unit of the bundle, which is a biologically active organelle. This project is in collaboration with M.F. Schmid, Maya Topf, A. Sali and P. Matsudaira.

Advances in single particle cryoEM have recently made it possible to reconstruct images of macromolecular machines to subnanometer resolutions. At this resolution range, a significant challenge arises in the analysis and annotation of the density map. To this end, we have developed a software toolkit (AIRS) that provides a robust, user-friendly environment for deciphering the structure of the macromolecular components (M. L. Baker, T. Ju and W. Chiu Structure 15:7-19, 2006). Application of some of these novel tools has resulted in the first de novo backbone trace in a single particle cryoEM reconstruction (4Å resolution map of unliganded GroEL provided by Steve Ludtke and Donghua Chen). Analysis of this model against a GroEL X-ray structure validated the accuracy of the cryoEM based model and has demonstrated the feasibility of building models in intermediate resolution cryoEM density maps.

Electron cryo-tomography is used to reconstruct unique molecular assemblies or one cell at a time through a tilt series. Due to the limited dose and tilt angles, it requires extensive data processing with multiple tomograms to extract, align and average components within a large assembly in order to visualize more structural details. We have introduced tomohunter software for this purpose. We will illustrate its application to herpesvirus capsid to allow the detection and visualization of a single portal vertex in a 125nm wide capsid particle (Chang, J., M. F. Schmid, F. Rixon, and W. Chiu J Virol, 81: 2065-2068, 2007).

With electron cryo-tomography, full three-dimensional models can be computed without the artifacts and difficulty of sectioning. Human blood platelets are a uniquely valuable specimen because they exhibit many of the hallmarks of eukaryotic cells, but are small enough that frozen-hydrated cells can be examined easily with tomography. We present reconstructions of platelets which provide a fresh, immersive look at eukaryotic fine structure. This project is in collaboration with J Chang and J Lopez.

NCMI is a Center supported by the National Center for Research Resources of NIH. Our Center focuses on the technology and research development for extending the resolution, speed, and flexibility of cryoEM for 3-D structure determination of biological nano-machines and cells. The resource tackles structural problems that are inappropriate or impossible for X-ray crystallography, NMR spectroscopy or light microscopy. Our Center develops various image processing and structure analysis software such as EMAN, SAVR and AIRS, which are distributed free of charge (http://ncmi.bcm.edu/software). To apply for access to our facility, see http://ncmi.bcm.edu/ncmi/. Our Center also sponsors workshops and symposia on a regular basis to disseminate its imaging and data processing technologies to a broad scientific community.

Continuity is a computational tool for continuum problems in bioengineering and physiology, especially those related to cardiac mechanics and electrocardiology research. Python is employed for user interfaces, communication, object-oriented component integration and wrapping of computationally efficient FORTRAN functions. For computationally extensive simulations, Continuity employs a Message Passing Interface (MPI) Python module, MYMPI, for parallel programming of Python. This module allows communication between a root Python process and multiple FORTRAN processes that perform compute-intensive ODE integration over the order of tens or hundreds of thousands of time steps. Simulating the spread of electrical activation in a 3D finite element model of part of the rabbit heart (1024 elements, 1377 nodes, and 8192 gauss points) resulted in a maximum speedup over the serial code of 15x on 32 nodes. We expect the parallel efficiency to improve with larger problems. The use of the MYMPI module has made possible investigations into biological phenomena that take place of timescales exceeding one beat, e.g. the effect of sympathetic nervous system activation. http://www.continuity.ucsd.edu

In order to deliver cyberinfrastructure to the general scientific and biomedical research community, transparent access and ease of use is of critical importance. Applications in systematic modeling of biological processes across scales of time and length demand more and more sophisticated algorithms and larger and longer simulations. The increased level of sophistication requires that cyberinfrastructure developers either work closely with the applications scientists, or develop middleware that flattens the learning curve for these scientists to use the grid willingly and transparently. The need to develop generalized and reusable components for cyberinfrastructure developers and the desire for customized solutions by application scientists create a dilemma that may take time to resolve. One reasonable approach is to let each do what they know best, and bridge the gap through innovative research and technology development. In this approach, we adopt technology that enables applications users to execute applications in the grid environment without modifications, and without knowledge of specific computational resources being utilized. This separation of scientific application development and the subsequent use of the grid, means that the cost of entry to the grid is minimal. Here we demonstrate the latest advances in the use of Gfarm-FUSE as a computational data grid, with CSF4 as the metascheduler, through a GridSphere portal based environment, termed My WorkSphere. The design and performance of this transparent grid computing environment will be demonstrated using MEME as an example. All the components developed or utilized are open source and available freely. https://nbcr.net:8443/worksphere/start.

Vision is a visual-programming environment in which a user can interactively build networks describing novel combinations of computational methods, and yielding new visualizations of their data without actually writing code. Nodes can be defined or modified interactively during a session and be saved. Multiple networks can be loaded at the same time. Sub-networks can be encapsulated in Macro nodes. Many of the nodes available in the Vision's library expose the functionality of the same packages used to develop Python Molecular Viewer (PMV). PMV has most of the features usually expected in a molecule viewer, but is dynamically extensible, i.e., new commands can be developed independently and placed in libraries. Python Molecular Viewer (PMV) is used together with Visual Programming Environment - Vision - to demonstrate how to run and analyze results from Adaptive Poisson-Boltzmann Solver (APBS) and a number of Web services such as PDB2PQR, Babel, MSMS and VIPERdb. Graphical User Interface (GUI) provided by PMV is a convenient way for setting up and calculating electrostatic potential, binding and salvation energies for biologically relevant molecules using APBS. The user is able to display 3D structure of the molecule and analyze the results by mapping electrostatic potential onto molecular surface and/or displaying different isocontours. It is also possible to run APBS jobs remotely using secured Web service. http://mgltools.scripps.edu

GEMSTONE is a rich client interface to an important set of grid-enabled computational chemistry and biochemistry tools. It incorporates a full end-to-end web services architecture for grid computing, including data management, remote job creation, and access to computational applications. Gemstone can run on any platform that is supported by the Firefox web browser. Among the applications it supports are APBS, GAMESS, and AUTODOCK. GEMSTONE is being developed with support from the NSF National Middleware Initiative, and through collaborations between the San Diego Supercomputer Center (SDSC) at UCSD; the National Biomedical Computation Resource (NBCR) at UCSD; the University of Zurich; the University of Texas, El Paso, and the Center for Theoretical Biological Physics. http://gemstone.mozdev.org.

The Grid-based computational infrastructure enables large-scale scientific applications to be run on distributed resources and coupled in innovative ways. However, users have to learn how to generate security credentials, stage inputs and outputs, access Grid- based schedulers, and install complex client softwares to take advantage of it. Scientific applications wrapped as Web services alleviate some of these problems by hiding the complexities of the back-end security and computational infrastructure, only exposing a simple SOAP API that can be accessed programmatically by application- specific user interfaces. We have developed Opal, a toolkit for wrapping scientific applications as Web services. Opal provides features such as scheduling, standards-based Grid security, and data management in an easy-to-use and configurable manner. We will present some of the computational chemistry and bioinformatics applications that have been deployed using Opal, and demonstrate the steps involved. Opal- based application Web services can be accessed by a multitude of clients, with GSI authentication if need be. They may also be used to compose complex scientific workflows, and we will describe the use of clients such as Kepler and Gemstone for the same. Finally, we will present some of our current and future work which includes integration with WSRF, meta-scheduling using CSF4, and a new and improved software stack using technologies such as Axis2, Hibernate, and DRMAA. http://nbcr.net/services.

In many biomolecular systems, the calculation of binding constants is very important to understand the kinetics of protein ligand, or protein protein interactions. The modeling of diffusional processes is important in determining the binding constants using either discrete (particle-based, stochastic) or continuum (probalistic) methods. While both types of methods are widely used, Continuum methods are computationally less expensive and may include such phenomena as fluid dynamics and biomechanical deformations. Song et al published a finite element based algorithm for solving the steady-state Smoluchowski equation (SSSE) for continuum modeling of diffusion in the Biophysical Journal in 2004. Using the Finite Element toolkit (FEtk), the authors are able to develop realistic biomolecular geometries through adaptive meshing techniques, and a solver of the SSSE. Physiologically, the ligands and enzymes are often in a non-steady state condition, where the concentration and binding constants may be time-dependent. Cheng et al has just published a paper on the solution of the time-dependent Smoluchowski equation (TDSE) to model the mouse Acetylcholinesterase (AchE) system, which is a diffusion-limited reaction with plenty of real experimental data for validation of models. The biomolecular mesh representation was generated using the Levelset Boundary Interior Exterior Mesher (LBIE). The TDSE caculations depend on the FEtk and APBS for electrostatistic calculations and for iterative mesh refinement for better accuracy in the binding constant calculation. TDSE is able to solve both steady-state and time-dependent problems very efficiently, and the results show good agreement with experimental observations. The figure on the left shows the diffusion of acetylcholine from the edge (high conecntration) to the center (low concentration) and its consumption by enzymes over a 15 microsecond time period. This research establishes the foundation for integration of molecular scale studies into larger simulations at the cellular scale such as the neuromuscular junction, where AchE plays a critical role in nerotransmitter signaling pathways. http://mccammon.ucsd.edu/smol/.

A tight coupling between cell structure, ionic fluxes and intracellular Ca2+ transients underlies the regulation of cardiac muscle function. Alterations in the cell geometry, in the Ca2+ protein distribution and pathways involved in these coupled processes are now recognized to be primary mechanisms of cardiac dysfunction. We developed 3D continuum models in cardiac muscle cells to investigate how the distribution of membrane Ca2+ proteins (L-type Ca2+ channel clusters, Na+/Ca2+ exchanger, membrane Ca2+ leak) might affect Ca2+ signals in terms of amplitude, time-course and spatial features when the sarcoplasmic reticulum is inhibited pharmacologically. The model predictions are in qualitative agreement with published experimental data. A parallel finite element software package is developed to solve the PDF model equation systems in a timely fashion on cluster of computers.

The pandemic threat of the Avian Flu and other infectious diseases require the development of sophisticated modeling tools to aid the discovery of therapeutic compounds. We've developed a prototype pipeline for the use of Relaxed Complex Molecular Dynamics in drug development, from preparation of protein structures, to selection of MD snapshots, to simulations of mutations using molecular modeling techniques. The MD techniques will also be developed to perform rescoring of docking experiments to refine the selection of top hits from virtual screening experiments using AutoDock and hierarchical screening procedures. The large computational requirements for these studies demand the use of supercomputers such as the BlueGene, as well as distributed resources such as the Open Science Grid, TeraGrid and community resources such as the World Community Grid. The resulting software through the encapsulation of these new algorithms will be of greater use to a wide range of diseases and mechanistic studies of protein ligand interactions.

Cryo-electron tomography provides a unique window into native (frozen-hydrated) cells and organelles, and offers the opportunity to study the structure of biomolecular machines in their physiological environment. The low signal-to-noise and limited, anisotropic resolution of these tomograms present numerous challenges to macromolecular structure determination. We present our latest results with the ryanodine receptor (RyR), a main component of the calcium release unit in striated muscle, Using a 3-nm RyR structure, determined from cryo-EM analysis of isolated, detergent-solubilized receptors, as a reference for correlation searches, we are are able to locate RyRs in tomograms of isolated frozen-hydrated triad junctions, and compute an average 3D structure. At current resolution (~7 nm) the overall shape and handedness of the tetrameric RyR is discernable. Surprisingly, the patch of membrane surrounding the RyR has a concave shape, with a radius of curvature of approximately 10 nm, forming a kind of vestibule around the luminal opening of the RyR. This local bending of the membrane might be a consequence of the phospholipid acyl chains aligning parallel to the hydrophobic surface of the protein, which forms an oblique angle with the transmembrane axis of the receptor.

Cryo-electron tomography is a powerful, high-resolution technique that permits visualization of the three-dimensional structure of cells preserved in a close-to-native state. We are using this technique at the RVBC to characterize the comparative architecture of Prochlorococcus, an abundant and globally important marine cyanobacterium that contributes nearly half of the net primary production in certain regions of the open ocean. Prochlorococcus cells are preserved by rapid freezing and visualized using a 400-kV cryo-EM in zero-loss mode. Tomographic reconstructions of whole cells reveal that Prochlorococcus strains, which differ by less than 3% in their 16S rDNA sequences, have evolved significant differences in their cellular structure and organization when cultured under identical growth conditions. Tomography of frozen-hydrated cryo-ultramicrotome sections provide higher resolution cellular structures, enabling us to visualize distinct membrane-lined connections between the lumens of tightly-appressed intracytoplasmic membranes. By revealing the 3D architecture of microorganisms preserved in a close-to-native state, cryo-electron tomography is advancing our fundamental understanding of microbial physiology and niche differentiation.

Cryo-electron microscopy and tomography of frozen-hydrated cells and tissues is opening a new window on the organization of the cell and the operation of the molecular machinery inside. A major obstacle has been the preparation of specimens from bulk-frozen biological material that are suitably thin (a few hundred nanometers) for imaging in the transmission electron microscope. Cutting frozen-hydrated specimens with diamond knives yields irregularly shaped sections with compression and fracture artifacts. Through collaborations with the Harvard University Center for Nanoscale Systems and Albany NanoTech (SUNY), the RVBC is exploring the use of focused ion beams (FIB) to mill frozen-hydrated specimens for electron tomography. We have found that the ion beam does not detectably heat these very labile specimens, and structural detail in FIB-milled specimens such as bacteria is preserved artifact-free (Marko et al., Nature Methods, 2007, in press). Extension to other prokaryotic and eukaryotic cells, and to mammalian tissue, as well as to suspensions of isolated macromolecular suspensions, will be explored.

The RVBC aims at developing tools for the visualization of biological structures at different scales using cryo-EM techniques. The major computational tasks range from the development of algorithms for the 3D reconstruction (either of molecules in single-particle form or subcellular structures, by electron tomography) to tools for the interpretation of the resulting 3D volumes. The ribosome is an essential macromolecular machine responsible for the biosynthesis of proteins. It is globularly shaped and relatively stable in chemical composition, both qualities which make it an ideal subject for the cryo-EM single-particle reconstruction technique. By using this technique, we study the interactions of the ribosome with its factors at each step of translation (initiation, elongation, termination and recycling). Combined with sophisticated structural modeling and fitting techniques, we are addressing many questions regarding the fundamental structure and the dynamic function of this molecular machine. A great challenge in visualizing a dynamic system by single-particle reconstruction is the heterogeneity of the molecule population in the specimen. One of the core projects of the RVBC is devoted to the development of efficient classification methods. Both supervised and unsupervised classification approaches are being explored in the study of translation.

Cryo-electron tomography provides a unique window into native (frozen-hydrated) cells and organelles, and offers the opportunity to study the structure of biomolecular machines in their physiological environment. The low signal-to-noise and limited, anisotropic resolution of these tomograms present numerous challenges to macromolecular structure determination. We present our latest results with the ryanodine receptor (RyR), a main component of the calcium release unit in striated muscle, Using a 3-nm RyR structure, determined from cryo-EM analysis of isolated, detergent-solubilized receptors, as a reference for correlation searches, we are are able to locate RyRs in tomograms of isolated frozen-hydrated triad junctions, and compute an average 3D structure. At current resolution (~7 nm) the overall shape and handedness of the tetrameric RyR is discernable. Surprisingly, the patch of membrane surrounding the RyR has a concave shape, with a radius of curvature of approximately 10 nm, forming a kind of vestibule around the luminal opening of the RyR. This local bending of the membrane might be a consequence of the phospholipid acyl chains aligning parallel to the hydrophobic surface of the protein, which forms an oblique angle with the transmembrane axis of the receptor.

Many diseases as well as the complexity to treat certain diseases are associated with or are the direct result of malfunctioning or hyperactive ion channels or transporters. Conventional, real-time methods have been instrumental for characterization of normal and abnormal states of transport. However, they have not been useful for studying the vast majority of slow rate and no-current passing (electroneutral) transporters, as well as ion channels on cells that cannot be voltage clamped. It is for these reasons that we are exploring the use of extracellular electrochemical sensors for real time measurement of ion transport through channels and transporters. Ion-selective electrodes possess the speed and sensitivity to capture extracellular ion concentration changes from single channels. This enables noninvasive, functional characterization of ion channels and mapping of their location. Used in a modulation mode, the signal to noise ratio is increased in order to measure small, relatively steady fluxes from ion transporters. This is useful for real-time characterization of slow rate or electroneutral transporters. These approaches are useful for the noninvasive, long-term monitoring of ion transport in the normal and diseased states.