Simulation and Computation

Research Emphasis

The overall goal of the Center for Bioelectric Field Modeling, Simulation, and Visualization is to develop and disseminate new methods, algorithms, and software systems for use in the study of experimental, clinical, and computational bioelectric field problems.

Current Research

Develop and implement the following techniques for the efficient manipulation and processing of bioelectric field data: geometric model generation and manipulation, bioelectric field simulation, and scalar and vector field visualization. Use the resulting software modules in supporting research projects within the center and in combination with center collaborators in computational, clinical, and basic electrocardiology and electroencephalography. Develop, disseminate, and support BioPSE (biomedical problem solving environment), an integrated, extensible, computation workbench that features a computational steering framework for interactive modeling, simulating, and visualizing bioelectric field problems; and map3d, a surface-based visualization package for qualitative and quantitative visualization and interrogation of time-dependent bioelectric field data.

Resource Capabilities

The center conducts research on and disseminates state-of-the-art software for geometric modeling, simulation, and visualization in basic and clinical bioelectric field research. Specific software tools include:

Modeling tools: Semi-automatic segmentation, surface generation, automatic mesh generation, and model editing, manipulation, and re-sampling tools.

Simulation tools: Finite element, finite difference, and boundary element techniques for the numeric solution of bioelectric field problems. Regularization techniques to constrain the effects produced by the ill-posed nature of ECG and EEG inverse problems; currently supported techniques include Tikhonov and Greensite regularization methods. Future directions include admissible solution approaches, and adaptive refinement techniques for forward and inverse approximation methods.

Visualization tools: Interactive scalar field display; isocontour and isosurface extraction; volume and surface rendering; vector field visualization. Current initiatives include quantitative spatio-temporal visualization, and methods for the characterization, representation, and presentation of error and uncertainty due to modeling, simulation, and visualization. Efforts are also being applied toward remote visualization tools.

Research Emphasis

The resource’s mission is to develop new methods, optimize existing approaches, and undertake research projects in biomedical areas that require high-performance computing, broadly construed to include large-scale data management, high-speed networking, and visualization. The resource also identifies new biomedical application areas that could benefit from high-performance computing, and speed the introduction of high-performance computing techniques into these areas.

Current Research

Current efforts are in structural biology, bioinformatics, cellular microphysiology, neural modeling, the Visible Human Project, and pathology. Specific projects include development and application of algorithms for sequence-sequence, sequence-structure, and multiple sequence alignment; classification and analysis of gene and protein superfamilies; understanding divalent metal ion binding sites in proteins and nucleic acids; incorporation of polarization effects in simulations of biopolymers; simulation of neural transmission (Mcell); simulation of neural networks on parallel platforms (NEOSIM); analysis of multi-electrode recordings of brain activity; display of anatomic (visible human) images; image analysis of pathology slides; databasing and retrieval of medically relevant images.

Resource Capabilities

Hardware

PSC’s Terascale Computing System has a peak speed of 6 Teraflops, 3 Terabytes of physical memory, and almost 70 TB of disk space. Currently the most powerful computer in the nation dedicated to open scientific research, it consists of 750 HP/COMPAQ ES-45 nodes, each with 4 processors rated at 2 Gflops peak, 4 GB of memory, and over 50 GBs of local disk. PSC also offers a 512 processor Cray T3E. Each processor is a 450-MHz alpha processor, with a peak speed of 900 Mflops and 128 MB of memory. For bioinformatics, the resource provides a dedicated 4-processor minisupercomputer with 6 GB of memory and 158 GB of disk; 525 Mb/s of commodity connectivity, with connectivity to two leading-edge networks (Abilene and vbns); archival system that can currently store 165 terabytes without human intervention, based on two storage-Tek silos, and Cray Research J90 file server with 20 processors and more than 500 GB of disk.

Software

More than 350 packages in quantum chemistry, molecular modeling, and genetic sequencing. All major sequence and structural databases. Most commercial packages for fluid dynamics, structural analysis, finite element analysis, mathematics libraries, equation solvers, tools, and graphics.

Training Facility

28 Silicon Graphics Indy graphics workstations for teaching workshops including graphically intensive subjects.

1. Gomez, C. M., Maselli, R., et al., A novel delta subunit mutation in slow-channel syndrome causes severe weakness by novel mechanisms. Annals of Neurology 51:102–112, 2002.
2. Nicholas Jr., H. B., Ropelewski, A. J., and Deerfield II, D. W., Strategies for multiple sequence alignment. Biotechniques 32:592–603, 2002.
3. Wetzel, A. W., Pomerantz, S. M., et al., Distributed multiuser visualization of time varying anatomical data 30th AIPR workshop: Analysis and understanding of time varying imagery, Oct. 10–12, 2001, Washington DC, IEEE Computer Society Press, pp. 109–114.

Research Emphasis

The resource is developing a user-friendly software package, S.A.G.E., that may be used to analyze family data to determine if the variability of a trait, either quantitative or qualitative, is significantly due to Mendelian segregation at a single genetic locus; if there is association between a quantitative or qualitative trait and a known polymorphic genetic marker; and if there are genetic loci linked to a known genetic marker that underlie variability in a trait.

Current Research

Theoretical development of statistical methods for the analysis of family data, especially to detect and identify genetic components that underlie disease susceptibility. Incorporating these methods into appropriate computer programs and making the programs generally available to other human geneticists and genetic epidemiologists in a well-documented and user-friendly form. Testing the validity, power, and robustness of the statistical procedures, especially to differentiate genetic causes from alternative environmental causes for familial aggregation. Application of methods and programs in collaborative projects to identify single genes that play roles in the etiology of various diseases.

Resource Capabilities

Software

The Human Genetic Analysis Resource (HGAR) utilizes a wide array of operating systems including Digital UNIX, Microsoft Windows 95/98 and NT, Sun Solaris, IBM AIX, SGI IRIX, and Linux. For large-scale simulations and computations, MOSIX Cluster Management Software is used to distribute workloads over a dedicated CPU cluster. Software development at HGAR proceeds in a variety of languages including C, C++, Python, Java, Perl, PHP, Fortran, and LotusScript. HGAR staff have access to a wide variety of math and statistical packages including Mathematica, SAS, and SPlus, usually utilizing the high-speed server systems directly from their desktops natively and via X-Windows emulation. Secure, distributed medical and research databases are being developed using Lotus Domino, Oracle8, and MySQL.

1. Olson, J. M., Goddard, K. A. B., and Dudek, D. M., A second locus for very-late-onset Alzheimer disease: A genome scan reveals linkage to 20p and epistasis between 20p and the amyloid precursor protein region. American Journal of Human Genetics 71:154–161, 2002.
2. Keen, K. J. and Elston, R. C., A problem in ascertainment. Communications in Statistics 30:1615–1631, 2002.
3. Burton, P. R., Palmer, L. J., et al., Ascertainment adjustment: Where does it take us? American Journal of Human Genetics 67:1505–1514, 2001.
4. Buxbaum, S., Elston, R. C., Tishler, P. V., and Redline, S., Genetics of the apnea hypopnea index in Caucasians and African Americans. Genetic Epidemiology 22:243–2530, 2001.

Research Emphasis

Problems in structural biology increasingly require researchers to move between models of low-resolution and detailed atomic models to fully explore and exploit experimental information. This resource focuses on development of new and integrated approaches to multiscale modeling, with an emphasis on modeling large-scale assemblies of nucleic acids and proteins with nucleic acids; developing methods that combine lattice-based dynamic Monte Carlo and all atom molecular dynamics; studying physical processes involved in and developing models for the interactions associated with virus assembly; and establishing new tools for the combined treatment of crystallographic and low-resolution structural models from cryo-electron microscopy. These research threads are tied together through the development and distribution of computer codes to make such multiscale simulations and modeling readily accessible to the scientific community at large.

Current Research

Modeling very large conformational changes occurring in proteins, nucleic acids, and their assemblies; developing methods and models to explore virus swelling and associated large-scale capsid dynamics during viral maturation; exploring of meso-scale distortions of molecular assemblies using low-resolution data from electron microscopy, in the absence of any atomic level structural information; providing links between low-resolution images of functional states of the ribosome during translocation and the near-atomic structural distortions that comprise these motions; characterization of protein-protein interfaces in assembled virus capsids from an energetic and structural standpoint, providing a basis for understanding large-scale molecular assembly. Ongoing development of methods for, and applications to, protein folding, loop, and homology modeling, including participation in CASP5, to perfect and “harden” physics-based approaches to structural genomics. Develop and test software to extend the range of atom-based modeling methods to larger systems.

Resource Capabilities

This resource is equipped with high-performance parallel Linux clusters of 32 and 64 processors as well as several dual-processor Silicon Graphics graphics servers. Large-scale modeling and simulation studies are performed on TSRI servers, which include a 256-node SGI Origin cluster.

Software under development includes lattice-based Monte Carlo sampling codes, nab (a software package to rapidly construct nucleic acid structures at atomic resolution), yammp (a molecular mechanics and modeling code directed toward low-resolution modeling of RNA and DNA), SITUS (for multiscale modeling of atomic and cryo-electron microscopy structural models), X-ray visualization and refinement software, and modules for CHARMM and AMBER. Much of this software is integrated for large-scale “ensemble” modeling, as relevant to structural genomic efforts, through the MMTSB tool set. The tool set, a suite of Perl libraries and routines that integrate and control the execution and management of large molecular simulations, is available at the MMTSB web site (http://mmtsb.scripps.edu/).

1. Tama, F. and Brooks, C. L. III, Mechanism and pathway of pH induced swelling in cowpea chlorotic mottle virus. Journal of Molecular Biology 318:733–747, 2002.
2. Tama, F., Brooks, C. L. III, and Wriggers, W., Exploring global distortions of biological macromolecules and assemblies from low-resolution structural information and elastic network theory. Journal of Molecular Biology 321:297–305, 2002.
3. Fiser, A., Feig, M., Brooks, C. L. III, and Sali, A., Evolution and physics in comparative protein structure modeling. Accounts of Chemical Research 35:413–421, 2002.
4. Skolnick, J. and Kolinski, A., A unified approach to the prediction of protein structure and function. Advances in Chemical Physics 120:131–192, 2002.

Research Emphasis

The mission of the National Biomedical Computation Resource (NBCR) at the University of California, San Diego (UCSD) is to conduct, catalyze, and enable biomedical research by harnessing advanced computational technology. To fulfill this mission, NBCR efforts are focused on four key activities: integrate computational and visualization tools in a transparent, advanced computing environment to enhance access to distributed data, computational resources, and instruments; develop and deploy advanced computational tools for modeling, data query, linking of data resources, 3-D image processing, and interactive visualization; provide access to and support of advanced computational infrastructure for biomedical researchers; and train a cadre of new researchers to have interdisciplinary knowledge of biology and the latest computation technologies.

The ultimate goal of the resource is to facilitate biomedical research by providing access to advanced computational and data grid capabilities via easy-to-use web portals, thereby enabling researchers to focus on the essential aspects of the biological/biomedical problem.

Current Research

NBCR is part of UCSD’s Center for Research on Biological Structure, and its technology development activities involve collaborations among researchers at UCSD, the San Diego Supercomputer Center (SDSC), the California Institute of Telecommunications and Information Technology (Cal-(IT)²), and The Scripps Research Institute, with a general interest in performing basic biomedical research from atomic to organismic levels. Core research projects include methods for pattern recognition in protein and nucleic acid structure, parallel tomographic methods for reconstruction of 3-D images, distributed database for cell-centered data, development/enhancement of cardiac electromechanics, parallel quantum mechanical modeling methods including environmental effects, development of platform-independent visualization tools, and the creation of portals for the biomedical community.

Resource Capabilities

NBCR provides web portals to a variety of analyses performed on its high-performance computing systems and servers. Software and services (see web site) include MEME, MAST, MetaMEME and SeqWeb, EULER, CE, MSMS, MIA, CMSMBR, GAMESS, QMView, PVM.

1. Yerushaimi, R., Noy, D., Baldridge, K., and Scherz, A., A mutual control of axial and equatorial ligands: Model studies with [Ni]-bacteriochlorophyll-a. Journal of American Chemical Society 124:8406–8415, 2002.
2. Martone, M. E., Gupta, A., et al., A cell-centered database for electron tomographic data. Journal of Structural Biology 1-2:145–155, 2002.
3. Sanner, M. F. and Olson, A., ViPEr, a visual programming environment for python. Proceedings of the 10th International Python Conference, Alexandria VA, Feb 4–7, 2002.
4. Baker, N. A., Sept, D., et al., Electrostatics of nanosystems: Application to microtubules and the ribosome. Proceedings of the National Academy of Sciences USA 98:10037–10041, 2001.
5. McCulloch, A. D., Sung, D., et al., Computational and experimental modeling of ventricular electromechanical interactions. In N. Virag, O. Blanc, and L. Kappenberger, Eds. Computer Simulation and Experimental Assessment of Cardiac Electrophysiology (pp. 89–96). Armonk, NY: Futura Publishing, 2001.
6. Reddy, B. V. B., Li, W. W., et al., Conserved key amino acid positions (CKAAPs) derived from analysis of common substructures in proteins. Proteins 42:148–163, 2001.
7. Reiter, L. T., Potocki, L., Chien, S., Gribskov, M., and Bier, E., A systematic analysis of human disease-associated gene sequences in Drosophila melanogaster. Genome Research 11:114–125, 2001.

Research Emphasis

The National Resource for Cell Analysis and Modeling (NRCAM) is developing methods for modeling cell physiological processes in the context of the actual 3-D structure of individual cells. Approaches in computational cell biology are coupled with high-resolution light microscopy to facilitate the interplay between experimental manipulation and computational simulation of specific cellular processes.

Current Research

NRCAM is developing the Virtual Cell, a general computational framework for modeling cell biological processes. This new technology associates biochemical and electrophysiological data describing individual reactions with experimental microscopic image data that describes their subcellular locations. Individual processes are integrated within a physical and computational infrastructure that will accommodate any molecular mechanism. Current development of the Virtual Cell is focused on expanding the generalized mathematical descriptions to include additional cell biological mechanisms, enhancing accessibility to biologists studying different biological processes, and integrating the interface with a database of images and reaction mechanisms. Current applications of the Virtual Cell include studies of calcium dynamics in neuroblastoma cells and Purkinje cells, and studies of intracellular RNA trafficking in oligodendrocytes. Additional collaborative research projects include modeling diffusional processes in mitochondria, nuclear transport, and aspects of cell motility.

Resource Capabilities

Microscopy instrumentation includes 3 confocal laser scanning microscopes including UV excitation, nonlinear optical microscopy utilizing a titanium sapphire pulsed laser, confocal-based fluorescence correlation spectroscopy, wide-field imaging workstation with cooled CCD and rapid excitation filter wheel, and dual-wavelength spectrofluorometer. Access to the facilities and technical staff is open to all researchers.

Computational resources include a Compaq Alpha cluster of 8 dual-processor DS20s, 2 RAID fileservers, 2 database servers and 2 primary servers, and multiple PC and Linux platforms running a variety of imaging processing and development software.

Authorized access to the Virtual Cell Modeling software is available via the Internet through a JAVA-based interface (www.nrcam.uchc.edu).

1. Slepchenko, B. M., Schaff, J. C., Carson, J. H., and Loew, L. M., Computational cell biology: Spatiotemporal simulation of cellular events. Annual Review of Biophysics and Biomolecular Structure 31:423–441, 2002.
2. Smith, A. E., Slepchenko, B. M., Schaff, J. C., Loew, L. M., and Macara, I. G., Systems analysis of Ran transport. Science 295:488–491, 2002.
3. Schaff, J. C., Slepchenko, B. M., and Loew, L. M., Physiological modeling with the Virtual Cell framework. Methods in Enzymology 321:1–23, 2000.

Research Emphasis

The objective of this multicenter resource (Beth Israel Deaconess Medical Center/Harvard Medical School; Massachusetts Institute of Technology, Division of Health Sciences and Technology; Boston University, Center for Polymer Studies; and McGill University, Department of Physiology) is to accelerate current research progress and catalyze new investigations in the quantitative study of complex physiologic signals. The resource has three interdependent components: PhysioBank is a large and growing archive of well-characterized digital recordings of physiologic signals and related data for use by the biomedical research community. PhysioBank currently includes databases of multiparameter cardiopulmonary, neural, and other signals from healthy subjects and from patients with a variety of conditions with major public health implications, including life-threatening arrhythmias, sleep apnea, neurologic disorders, and aging. PhysioToolkit is a library of open-source software for physiologic signal processing and analysis, and detection of physiologically significant events using both classical techniques and novel methods based on statistical physics and nonlinear dynamics. PhysioNet is an online forum for dissemination and exchange of recorded biomedical signals and open-source software for analyzing them.

Current Research

The resource is developing new algorithms that quantify the transient and local properties of nonstationary physiologic signals and the cross-interactions among multiparameter signals. These techniques will be used to detect changes that may precede the onset of catastrophic physiologic events, including epilepsy and sudden cardiac death. Complementary studies are aimed at developing techniques to quantify the nonlinear dynamics of physiologic control, with an emphasis on modeling these mechanisms and identifying new measures that have diagnostic/prognostic utility in life-threatening cardiopulmonary pathologies, such as sleep apnea and congestive heart failure. Another core area of research is the development of methods for assessing signal quality in multiparameter data.

Resource Capabilities

The PhysioNet web site creates an online community for the dissemination and exchange of recorded biomedical signals and the software for analyzing them by providing facilities for cooperative analysis of data and evaluation of proposed new algorithms. Much of the PhysioBank and PhysioToolkit software utilizes standard networking protocols, allowing interactive display and analysis of physiologic signals at remote locations on the Internet.

1. Goldberger, A. L., Amaral, L. A. N., et al., Fractal dynamics in physiology: Alterations with disease and aging. Proceedings of the National Academy of Sciences USA 99[suppl 1]:2466–2472, 2002.
2. Bub, G., Shrier, A., and Glass, L., Spiral wave generation in heterogeneous excitable media. Physical Review Letters 88:058101, 2002.
3. Costa, M., Goldberger, A. L., and Peng, C. K., Multiscale entropy analysis of complex physiologic time series. Physical Review Letters 89:068102, 2002.
4. Moody, G. B., Mark, R. G., and Goldberger, A. L., PhysioNet: A web-based resource for the study of physiologic signals. IEEE Engineering in Medicine and Biology 20:70–75, 2001.
5. Amaral, L. A. N., Ivanov, P. Ch., Aoyagi, N., Hidaka, I., Tomono, S., Goldberger, A. L., Stanley, H. E., and Yamamoto, Y., Behavioral-independent features of complex heartbeat dynamics. Physical Review Letters 86:6026–6029, 2001.

Research Emphasis

The Resource for Biocomputing, Visualization, and Informatics (RBVI) creates innovative computational and visualization-based data analysis methods and algorithms; implements these as professional-quality, easy-to-use software tools; and applies these tools for solving a wide range of genomic and molecular recognition problems within the complex sequence-structure-function triad. Application areas include gene characterization and interpretation, drug design, variation in drug response due to genetic factors, protein engineering, biomaterials design, and prediction of function from sequence and structure.

Current Research

Sequence analysis and bioinformatics: The characterization and interpretation of genomic data, including knowledge discovery and transfer in nucleic acid and protein sequence analysis, pharmacogenomics, identifying gene and regulatory motifs, protein family/superfamily relationships, and gene expression patterns.

Structural informatics: The development, application, and dissemination of analysis methodologies and software tools in computational structural biology, including algorithm development for low- and high-resolution protein structural models and their comparison, molecular visualization for structural analysis and the integration of sequence and tertiary structural information, and facilitation of collaborative research, especially with distant scientists, through use of high-performance network technology.

Functional informatics: Theoretical and applied research in how protein structures deliver function, including the identification and characterization of protein superfamilies, the generation of new computational-based representations of protein chemistry, and the development of structurally contextual definitions of protein function.

Resource Capabilities

Hardware

High-performance cluster of symmetric multiprocessor Hewlett-Packard AlphaServer computers for performing theoretical studies on protein and nucleic acid structure and function, and for storing, searching, and analyzing various sequence and structure databases. High-performance interactive three-dimensional graphics workstations equipped with special glasses for viewing in stereo for visualization of complex molecular structures. All systems are interconnected via a high-performance network and are capable of distributed computations.

Software

Commercial applications for database searching and analysis, and several locally developed packages disseminated as documented source code to allow others to use the resource’s software both for their own research applications and as a starting point and training tool for specialized applications. Local packages include MidasPlus, a molecular visualization application used to display and interactively manipulate macromolecules such as proteins and nucleic acids, and Chimera, an advanced molecular visualization system that can easily be extended for specialized molecular modeling needs.

1. Stryke, D., Huang, C. C., et al., SNP analysis and presentation in the pharmacogenetics of membrane transporters project. In Pacific Symposium on Biocomputing 2003 (R. B. Altman, A. K. Dunker, L. Hunter, K. Lauderdale, and T. E. Klein, eds.). Singapore: World Scientific Publishing, December 2002.
2. Konerding, D. E., Huang, C. C., and Ferrin, T. E., Chimera: Affordable desktop molecular modeling on Linux workstations. Proceedings of the 5th Annual Linux Showcase and Conference, Usenix Association, November 2001. Available from www.cgl.ucsf.edu/home/tef/pubs/konerding.pdf.

Research Emphasis

The resource studies large biomolecular processes in living cells, focusing on membrane proteins that mediate the exchange of materials and information across, in particular, biological membranes as well as the conversion between electro-osmotic, mechanical, and chemical energy. It also develops software for large-scale simulations. Software tools include NAMD, a molecular dynamics simulation program used for classical, atomistic molecular dynamics simulations of large biomolecular aggregates; VMD, a molecular visualization program for displaying, animating, and analyzing both large and small biomolecular systems using 3-D graphics and built-in scripting; BioCoRE, a web-based, tool-oriented collaboratory for biomedical research and training.

Current Research

Interactive molecular dynamics (IMD) for the manipulation of molecular simulations with real-time force feedback and interactive display; investigations of aquaporin channels, mechanosensitive channel, ATP synthase, chloride channel, photosynthetic proteins, visual receptors, and proteins with mechanical functions; efficient evaluation of force fields and integration schemes for simulation of very large biomolecular systems; efficient distributed molecular dynamics programs on workstation clusters and massively parallel machines; continued development of NAMD, VMD, and BioCoRE.

Resource Capabilities

Instruments

Three main computational platforms: 100 Athlon PC nodes running as four Scyld Beowulf Linux clusters, the primary development platform for the group’s molecular dynamics program, NAMD; 4 Microway Alpha AXP21264-500 dual-processor nodes, two with 4 GB of memory, primarily used for large-memory jobs such as quantum simulations; and 2 dual-processor SunBlade 2000 systems with XVR1000 video cards and 4 GB of memory, used for both visualization and large memory jobs. These systems are controlled using a shared queueing system. Four Sun Enterprise 250 servers with a combined total of 2.75 TB of disk space serve the network-wide home and project directories. The servers are backed up daily on four DLT4000 tape changers. The key graphics platforms include aforementioned SunBlade 2000 systems; one Sun Ultra 80 and one Ultra 60 system, each with Expert 3-D graphics; and an eight-processor SGI Onyx2 with InfiniteReality graphics. Nearly all researcher workstations are equipped with 3-D video cards and are capable of real-time stereoscopic visualization. The Sun Ultra 80 drives a stereo-capable Electrohome Marquee 8500LC projector in the resource’s 3-D projection facility.

Special Features

IMD, interactive molecular dynamics; VMD, a visualization program for interactive display and animation of molecules; NAMD, a parallel message-driven molecular dynamics program reaching teraflop performance; and BioCoRE, an integrated set of computational tools that functions as an interactive visual computing environment for the simulation and collaborative study of biopolymers over distance and as a training platform.

1. Phillips, J. C., Zheng, G., Kumar, S., and Kale, L. V., NAMD: Biomolecular simulation on thousands of processors. Proceedings of the IEEE/ACM SC2002 Conference at http://dlib.computer.org/conferen/sc/1524/pdf/15240036.pdf.
2. Tajkhorshid, E., Nollert, P., et al., Control of the selectivity of the aquaporin water channel family by global orientational tuning. Science 296:525–530, 2002.
3. Jensen, M. O., Park, S., Tajkhorshid, E., and Schulten, K., Energetics of glycerol conduction through aquaglyceroporin GlpF. Proceedings of the National Academy of Sciences USA 99:6731–6736, 2002.