Speaker: Dr. Gerhard Klimeck, Professor and Technical Director, NSF - Network for Computational Nanotechnology (NCN), Purdue University, W. Lafayette, IN

Abstract. The National Science Foundation envisions future cyberinfrastructures to be built on the four pillars of virtual organizations, high performance computing, visualization and data.  nanoHUB.org embodies such a cyberinfrastructure today that serves over 24,000 users worldwide on an annual basis.  Over 5,500 of these users have run over 217,000 simulations, including HPC-based runs and visualized date interactively. nanoHUB is used by all top 50 engineering schools.  Over 50 applications are available online ranging from toy models to sophisticated simulation engines. The nanoHUB provides the resources for modeling, simulating and computing without any software installation to users with access to a web browser.  All nanoHUB services are freely open to the public. The nanoHUB represents a community–based, service–oriented, science architecture that leads the way for new cyberinfrastructure-enabled science. We have identified 81 citations in the literature by non-nanoHUB affiliated scientists and engineers.

This presentation will provide a nanoHUB overview and demonstrate how effective, user-friendly cyberinfrastructure can bridge education and research seamlessly for computational experts, experimentalists, educators, and students.  The development and deployment of the nanoelectronic Modeling Tool (NEMO 3-D) in the nanoHUB with its capability to perform multi-million atom electronic structure calculations on high-end HPC platforms as well as rapid second-fast executions in an educational module will be highlighted.  The presentation will conclude with a vision for future expansions into other HUBs for other fields.

Biosketch: Gerhard Klimeck is the Technical Director of the Network for Computational Nanotechnology at Purdue University and a Professor of Electrical and Computer Engineering since Dec. 2003. He was the Technical Group Supervisor of the High Performance Computing Group and a Principal Scientist at the NASA Jet Propulsion Laboratory. His research interest is in the modeling of nanoelectronic devices, parallel cluster computing, and genetic algorithms. Gerhard developed the Nanoelectronic Modeling tool (NEMO 3-D) for multimillion atom simulations. Previously he was a member of technical staff at the Central Research Lab of Texas Instruments where he served as manager and principal architect of the Nanoelectronic Modeling (NEMO 1-D) program. Dr. Klimeck received his Ph.D. in 1994 from Purdue University and his German electrical engineering degree in 1990 from Ruhr-University Bochum. Dr. Klimeck's work is documented in over 170 peer-reviewed publications and over 270 conference presentations. He is a senior member of IEEE and member of APS, HKN and TBP.

10: 45AM-11:00AM: Coffee Break

11:00AM-12:00PM (Room 1505 SC, Engineering Building's Large Auditorium): Invited Keynote Talk:

Biomedical Informatics: Overview and Collaborative Opportunities

Speaker: Dr. Brian D. Athey, Associate Professor of Biomedical Informatics, and PI, NIH- National Center for Integrative Biomedical Informatics (NCIBI), one of seven National Centers for Biomedical Computing (NCBC).

Biosketch: Dr. Brian Athey received his Ph.D. in Cellular and Molecular Biology (Biophysics concentration) from the University of Michigan (1990), with a research focus in macromolecular structural biology. Dr. Athey is currently an Associate Professor of Biomedical Informatics in the University of Michigan Medical School Department of Psychiatry; Brian is also an Associate Professor of Bioinformatics and Computational Biology in the UM Bioinformatics Program and he is an active course master in the Program. In addition, he is Director of the Michigan Center for Biological Information (MCBI), the Bioinformatics Computing and Digital Library Hub of the Michigan Life Science Corridor (MLSC) Infrastructure Initiative, administratively housed in the University of Michigan Office of the Vice President for Research (http://www.ctaalliance.org/MCBI/index.htm). Brian was recently named Associate Director of the University of Michigan Center for Computational Medicine and Biology (ccmb.med.umich.edu), where his is responsible for the computing and data infrastructure build out for the Center and the Medical School.

During his Ph.D. thesis research in the 1980s, Brian proposed the double helical crossed-linker model for the structure of chromatin, at that time controversial, but now thought to be the correct model for this critical macromolecular complex, and featured in many textbooks. In the mid 1990s, Brian served as the Director of Biological Imaging Programs at the Environmental Research Institute of Michigan (ERIM) (now part of General Dynamics). Dr. Athey is also a founding member of the Michigan Nanotechnology Institute for Medicine and Biology (www.nano.med.umich.edu). He was also a longtime collaborator and friend for the world famous and late Professor Emmett Leith, the inventor of off-axis holography. He and Leith invented and demonstrated an incoherent white light holographic microscope in 2001 with their student Kurt Mills and collaborator David Dilworth.

Brian is best known for his work with the National Library of Medicine (NLM) Visible Human Project where he has been a leader in establishing a nationwide Internet2 end-to-end test bed demonstration project with NIH/NLM sponsorship in collaboration with Pittsburgh Supercomputing Center, Stanford University, and the Uniformed Services University of the Health Sciences (USUHS) (vhp.med.umich.edu). This project naturally lead to the establishment of the very successful DARPA Virtual Soldier Project (www.virtualsoldier.us), a nationwide consortium which he runs that builds from the visible human project to extend from basic human anatomy to physiological modeling, functional simulation and prediction after a traumatic injury.

Brian is the Principal Investigator of the recently established (Fall, 2005) National Center for Integrative Biomedical Informatics (www.ncibi.org), only one of seven NIH National Centers for Biomedical Computing (NCBC), a center piece of the NIH Roadmap initiative. The mission of the National Center for Integrative Biomedical Informatics (NCIBI) is to facilitate scientific exploration of complex disease processes on a much larger scale than is currently feasible by developing interactively integrated analytical and modeling technologies to acquire or create context-appropriate molecular biology information from emerging experimental data, international genomic databases and the published literature.

Dr. Athey has published over 45 papers and book chapters in the scientific literature and has given well over 100 talks nationally, including several keynotes and plenary talks. Brian has been a Peace Fellow of the Federation of American Scientists (FAS) since 2000, and is active in international educational and treaty efforts to prevent biological warfare and terrorism. He is also active with the FAS Learning Federation (www.learningfederation.org). Dr. Athey is a founder and board of Advisor member of Scientist and Engineers for America (SEforA.org)

Parallel I/O and Storage Considerations for Performance of Scalable HPC Applications

Speaker: Stan Posey, Director, Applications and Industry Development, Panasas Inc, Fremont, CA, USA

Abstract: As expectations grow for Linux clusters to meet levels of capability for even challenge-scale HPC applications, the reality of overall scalability in practice is limited by I/O that will not scale on conventional file system implementations. As such, a new class of object-based parallel file systems and storage technology has developed that is designed to extend overall scalability of conventional clusters. These commodity storage systems can offer a large single resource of shared storage that provides parallel data access directly between cluster compute nodes and the storage system. In this way, a simulation has direct access to a storage system for both parallel computational I/O, and for subsequent post-processing and visualization of the results without the need for non-productive large file transfers. This approach eliminates the typical bottleneck of a head node collecting serial I/O streams from individual partitions, and also eliminates the need and expense of NFS servers. This technical presentation will examine the industry motivation for such parallel file systems and storage, and describe the features and benefits of a Panasas implementation, through examples of various HPC workflow requirements, for a variety of production-level applications.

Biosketch: Stan Posey currently manages the Panasas strategy of HPC Applications and Industry Development for a variety of target vertical markets. Mr. Posey is responsible for company guidance and consultation on Panasas advancement of HPC technology and solutions, with emphasis on applications in computational mechanics and sciences. Prior to joining Panasas in 2007, Mr. Posey has contributed for more than 20 years in various roles of HPC applications development including applications engineering and market development roles at SGI, research and engineering positions with the US Department of Energy, Oak Ridge National Laboratory, consulting and ISV software vendor CD-adapco Group, and systems vendor Control Data Corporation. Mr. Posey has co-authored several papers for HPC research journals and technical conferences, and his education includes a M.Sc. in Mechanical Engineering from the University of Tennessee, Knoxville, TN, USA.

Abstract: Orthology between gene pairs allows the identification of ancestral patterns of gene regulation. We have mapped the ortholgous gene pairs flanking bidirectional promoters, which regulate expression of protein coding genes. We then examine the evolutionary history of bidirectional promoters across multiple phyla. Using knowledge of the genes regulated by bidirectional promoters across long evolutionary distances, such as human to fish, we are able to predict when gene annotations are missing or where assembly problems exist in the genomic sequences of higher vertebrates. We use these orthology assignments to examine the evidence for selective pressure which maintains the bidirectional gene pairing versus a tail to tail arrangement of genes. Furthermore, we apply a discriminatory method, known as ESPERR, to test our ability to discriminate among two classes of promoters, bidirectional promoters and monodirectional promoters. The results of these analyses will lead to improved annotations of regulatory regions in vertebrate genomes and contribute toward the goal of de novo classification of promoters using sequence-based genomic features. This work was collaborating with Dr. James Taylor of NYU Courant Institute of Mathematical Sciences and Dr. Laura L. Elnitski, Head of Genomic Functional Analysis, National Genomics Research Institute (NHGRI), NIH.

Biosketch: Mary Qu Yang received her interdisciplinary Ph.D. degree in Biophysics and Computer Engineering with specialization in Computational Science and Engineering from Purdue University and holds Master of Electrical and Computer Engineering Degree in Computer Engineering, and Master of Science Degree in Biophysics, also both from Purdue. She was the recipient of an Outstanding Bilsland Dissertation Fellowship, the recipient of a NIH Post Doctoral Fellowship for National Human Genome Research and also the recipient of a NIH - Oak Ridge, DOE research specialist fellowship. She was trained as combined computational and experimental scientist with more than 15 years of teaching, research and engineering practice experience in the fields of software engineering and biomedicine. Dr. Yang has received a number of outstanding achievement awards, best paper awards and plenary keynote speaker invitations, and has been an editor of a number of journals and proceeding books. He is currently on the editorial boards of Journal of Supercomputing, International Journal of Bioinformatics Research and Applications, Journal of Computational Intelligence in Bioinformatics and International Journal of Data Mining and Bioinformatics. She has published more than 40 peer-reviewed papers and several book chapters.

Invited Talk: Developing Synergistic Bioinformatics and Machine Learning Approaches to Predict Potential Malignancies

Speaker: Dr. Jack Y. Yang, Harvard Medical School, Harvard University, Massachusetts, USA.