Over 100 people attended the fourth annual CI Forum on Oct. 22, 2015, traveling from around the Midwest and from a variety of backgrounds. The daylong event featured three keynote speakers, nine breakout lectures, and plenty of networking opportunities over breakfast, lunch, and an evening social hour. The goal — highlighting cyberinfrastructures to show the critical role they play in computational research — was a resonding success.
This year’s themes of information sharing echoed the world’s movement toward open data. For years, scientific research had shifted from the world stage to behind closed doors, but in recent years that trend is being reversed. From the standpoint of good science, sharing is once again being seen as prudent as well as ethical.
iCER Director Dr. Kennie Merz and MSU mathematics professor Dr. Andrew Christlieb shared the duty of providing the opening remarks, welcoming everyone to the Forum. They were immediately followed by Dr. John Towns, who provided the first keynote address, Supporting Science Communities with XSEDE. He talked about the role of XSEDE (Extreme Science and Engineering Discovery Environment) as a major research infrastructure, describing the collaborations being made by pan-disciplinary researchers from around the world. He also discussed how XSEDE intends to support supercomputing at the national level, and how it will advance the skills and capacity of the U.S. to utilize sophisticated modeling and simulation techniques so that it can address a wide range of problems, from manufacturing to the discovery of novel therapeutics.
The Morning Breakouts Sessions consisted of a live demonstration by Hossam Metwally from Ansys, Inc., entitled State of Art of Multiphysics Computational Simulation. MSU faculty members Hailey Mooney, Patricia Soranno, Joe Messina, Lynne Goldstein, James Pivarnik, and Connie Berg joined forces to present the benefits and challenges of open data in Sharing is Caring. And HPCC Director Sharan Kalwani and Administator Andrew Keen went back-to-back in their presentations, Current and In-Development HPCC Computational Resources (Keen) and Ongoing Developments and Advances in Future HPCC Technologies, (Kalwani).
After lunch, Dr. Craig Stewart gave the second keynote address, Cyberinfrastructure for Research. Among his titles, Dr. Stewart is the executive director of the Pervasive Technology Institute, an advanced IT research department at Indiana University. Craig gave a timely talk for iCER on how IU built up its supercomputing capacity to support both its local and national user base. He highlighted the role of hardware, and stressed how important it is to have relevant training and research to ensure the effective use of sophisticated compute hardware.
The Afternoon Breakout Sessions once again split the group up into smaller units. Dr. David Cerutti’s presentation, A Brief History of Molecular Dynamics, provided a concise analysis of how high-performance computing has played a crucial role this relatively young field. In the First Steps Towards Personalized Medicine, Dr. George Mias talked about how the burgeoning omics technologies are guiding efforts to make advances in the implementation of precision medicine. Another team-up of MSU researchers Greg Austic, Sven Böhm, Devin Higgins, and Graham Pierce presented Open Data at MSU. Dr. Tyce DeYoung talked about how high-performance computing is helping scientists learn about some of the smallest particles in the universe at the IceCube Neutrino Observatory in Antarctica in his presentation, Tracking Ghost Particles. And Dr. Brian O’Shea demonstrated how the HPC is helping study the life cycle of galaxies in The Universe in a Box.
The final keynote of the day was Data Access, Research Transparency and the Public Value of Science presented by Dr. Arthur Lupia, the Hal R. Varian Collegiate Professor of Political Science at U of M. Dr. Lupia presented a wide-ranging lecture on efforts toward broader sharing of primary data in the social sciences.
In his closing remarks, MSU’s Director of Libraries Clifford H. Haka acknowledged the university’s key collaborators, and re-emphasized the importance of information sharing in research. Dr. Corey Washington, the other closing speaker, offered his optimistic thoughts about researchers’ use of advanced cyberinfrastructures to increase paths to knowledge.
At the Networking Social Hour, the winners of the poster contest were announced. Tridip Das (Chemical Engineering) took the top prize with his poster, A Thermodynamics and First-Principles Based Approach to Develop Silver-Free Solid Oxide Fuel Cell Braze Alloys, which detailed a low-cost, clean energy conversion process. Greg Meece (Physics and Astronomy) and Parrissa R. Brown (Urban and Regional Planning) took second and third place, respectively.
The success of the 2015 CI Forum shows promise for the mainstream familiarity with cyberinfrastructures, as well as the shift back to a scientific world where data sharing is the norm, not the exception.
— The iCER Staff
What is CI?
The National Science Foundation defines CI as a collection of advanced technologies and services to support scientific inquiry. This includes:
- Computing clusters and high performance computing systems
- Data management, data integration, and data storage systems
- High speed networks
- Data mining and data visualization
- Collaboration and security tools
- The people who design, build, and run these systems
The CI Forum is coordinated by the Institute for Cyber Enabled Research (iCER). The CI Forum is jointly sponsored by the Office of the Vice President for Research and Graduate Studies, IT Services, and iCER. This event is based on an initiative originally supported by the National Science Foundation in 2010 to support the implementation and use of cyberinfrastructure at research institutions.
Sessions
Supporting Science Communities with XSEDE
Presented by Dr. John Towns
State of Art of Multiphysics Computational Simulation
Presented by Ansys
This presentation discussed industry current trends in the product development process. The topics of multi-disciplinary optimization, engineering democratization, and need for speed (high-performance computing) were also discussed.
Slide presentation (coming soon)
Sharing is Caring: Challenges and Opportunities in Producing Open Data
Presented by Hailey Mooney, Patricia Soranno, Joe Messina, Lynne Goldstein, James Pivarnik and Connie Berg
This session provided an overview of open data in academic research including a brief look at the history of U.S. federal government interest in open science data, the landscape of options for publishing data in the current system of scholarly communication, and library roles in support of data. Researchers shared viewpoints regarding arguments for sharing (or not) and look at disciplinary norms and trends around open data. The session also included practical considerations regarding research integrity and intellectual property. Discussion allowed speakers and the audience to discuss initiatives, both current and imagined for the future, on campus or in the broader academic community in support of data sharing.
- Slide presentation - Mooney
- Slide presentation - Soranno
- Slide presentation - Messina
- Slide presentation - Goldstein
- Slide presentation - Pivarnik
- Slide presentation - Berg
Current and In-Development HPCC Computational Resources
Presented by Andrew Keen
This presentation highlights the hardware resources the HPCC makes available to the MSU research communinty, including its current computational platforms, storage, and network capabilities. This presentation also discussed the new NSF-funded distributed storage project (OSIRIS) as well as a preview of a new cluster at MSU.
Ongoing Developments and Advances in Future HPCC Technologies
Presented by Sharan Kalwani
To further speed up research and accelerate discoveries in new research areas, there are a number of technologoies being developed. This talk highlights several projects, each of which can enable major improvements in areas including computational demand, larger problem sizes, make easier the job of expressing problems that can take advantage of newer architectures.
Cyberinfrastructure for Research: From Campus Growth to National Trends
Presented by Dr. Craig Stewart
A Brief History of Molecular Dynamics: An analysis of its place in high-performance computing
Presented by David Cerutti
First Steps Towards Personalized Medicine: Dynamic Omics Integration and MathlOmica
Presented by George Mias
The emergence and ready availability of novel - omincs technologies is guiding our efforts to make advances in the implementation of personalized (or precision) medicine. High quality genomic data is now complemented with other dynamic omes, (e.g. transcriptomes, preteomes, metabolomes, autoantibodyomes) and other data, providing temporal profiling of thousands of molecular components. Such information allows for network inference that can be leverages to personalized medicine through dynamically following the multiple omics technologies. Pathway changes over time and associated network interactions were first observed using multi-omic information from a proof-of-principle investigation, which explored the implementation of personalized medicine in an initially healthy individual over more than five years. Our investigataions extend this pilot to the study of diseases such as asthma and drug treatment. The analysis of dynamic omics data also necessitates developingnew statistical and computational methadology, towards the integration of the different platforms. We are continuously developing a framework, MathlOmica, to analyze mulitple omics and integrate temporal data, based on longitudinal experiments from our lab, with a focus on personalized medicine implementations. Our framework currently approaches transcriptome (RNA) sequencing, mass spectrometry (proteins and small molecules) and protein array data, and includes quantitation methods available for each analysis, and classification of all longitudinal omics. We hope that the combination of clinical studies and combined omics analyses will provide a medically relevant interpretation with potential applications to a more personalized, precise and preventative medicine.
Open Data at MSU
Presented by Greg Austic, Sven Bohm, Devin Higgins and Graham Pierce
This session shared examples of research projects with an open data element. Presenters focused on the mechanisms employed for data sharing, challenges encountered in the process, and how open data enriched the project.
- Slide presentation - Austic
- Slide presentation - Bohm
- Slide presentation - Higgins
- Slide presentation - Pierce
The Universe in a Box: Studying Galaxies on Supercomputers
Presented by Brian O’Shea
The study of structure formation in the universe, particular the formation and evolution of galaxies, is a problem that requires extensive use of large-scale computing. In this talk, the challenge of studying galaxy formation was discussed, as well as how that research is conducted using the world’s largest computers, highlighting the use of MSU's High Performance Computing Center (HPCC).
Tracking Ghost Particles: High-Performance Computing for the IceCube Neutrino Observatory
Presented by Tyce DeYoung
The IceCube Observatory is the world's largest detector of neutrinos, subatomic particles which interact so weakly with matter they have been dubbed "ghost particles." IceCube uses a cubic-kilometer section of the Antarctic ice cap below the U.S. South Pole station as a neutrino target, detecting light produced by the rare neutrinos which do interact with the ice. We use these neutrinos both to study the dynamics of some of the most energetic objects in the universe, and also to measure the fundamental properties of the neutrinos themselves. This talk discussed some of the computational challenges related to IceCube data analysis, and the utilization of high-performance computing to enhance the scientific output of the detector.
Data Access, Research Transparency and the Public Value of Science
Presented by Dr. Arthur Lupia
Quantitative social researchers have long been engaged in data sharing as exemplified by the founding of the Inter-university Consortium for Political and Social Research data archive in 1962. The American Political Science Association recently adopted new policies guiding data access and research. Since then, 27 journals have signed the Data Access & Research Transparency Joint Statement, committing to promote and implement policies requiring authors to make data as accessible as possible. However, creating open data still remains as an area of great challenge and opportunity.