2014 was a productive year for the Computation Institute, so much so that we can't even fit our entire year in review into a single post (as we did last year). From papers on influenza, climate modeling, whole genome analysis, and autism prevalence to workshops on research data management, urban research, and the science of science, the Computation Institute was active across the University of Chicago and Argonne National Laboratory campuses and beyond, facilitating collaboration between disciplines and applying powerful computational tools to accelerate the pace of science. In this post, we cover January through June of this busy 2014, with the second half of the year coming later this week. Happy Holidays!
A team of scientists at the Argonne’s Transportation Technology R&D Center, including CI Fellow Sibendu Som, used advanced computational modeling to simulate fuel injection and combustion in car engines, working with the automotive industry to develop more efficient engines and alternative fuels.
CI Director Ian Foster wrote about “invisible software” and “dark software,” highlighting the need for improved citing and publishing of software used in scientific research. He also produced a white paper on the subject for a Department of Energy workshop on software productivity for extreme-scale science.
Our inaugural Inside the Discovery Cloud Speaker Series kicked off with talks from Charlie Catlett on computational urban research and James Evans on metaknowledge.
The CI’s Hack Arts Lab, better known as HAL, celebrated its first birthday with an open house and demonstration of its digital fabrication technologies. The facility, directed by CI Fellow and Faculty Jason Salavon, provides the University of Chicago community with access to 3D printers, laser etchers, and advanced visualization tools.
Chicago played host to the annual American Association for the Advancement of Science meeting, where the Computation Institute organized several sessions on new developments in computational science. Ian Foster and CI Senior Fellow Folker Meyer spoke at a panel on “Outsourcing Science” with representatives from Science Exchange and nanoHUB, discussing the potential of these services for accelerating future research. CI Senior Fellow and Faculty Robert Grossman laid out his vision for a biomedical data commons (which would come to fruition in November), speaking with fellow researchers who have built systems for sharing data on drug research and cancer genomics. And CI Senior Fellow Charlie Catlett spoke alongside City of Chicago Chief Information Officer Brenna Berman and other researchers about the potential of data and computation to usher in a new wave of city research.
A paper from the laboratory of CI Senior Fellow Wim van Drongelen describes Kunstkop, their model of how epilepsy starts and spreads in a network of neurons. The development team, led by the CI’s Lorenzo Pesce, used the model on the University of Chicago biomedical supercomputer Beagle to run simulations involving hundreds of thousands of artificial neurons.
Beagle was also the platform for a barrier-shattering demonstration of genetic analysis, conducted by the cardiovascular genetics laboratory of Elizabeth McNally. Using the supercomputer, McNally’s team (which also included Pesce) analyzed 240 whole genomes in two days, easing the analytic bottleneck that has slowed the development of genetic medicine.
CI research center ARTFL, the Project for American and French Research on the Treasury of the French Language, received a National Endowment for the Humanities grant to create a digital “commonplace book,” mining data from 18th century texts and producing a new database for literature and history research.
Scientists at the Center for Structural Genomics of Infectious Diseases (CSGID) constructed APPLIED, a computational pipeline for finding new ligand partners for disease-related proteins. After 3D protein structures are determined by CSGID, the pipeline uses the supercomputing power of Argonne’s Mira to test tens of millions of ligands and find new interactions that could lead to more effective drugs.
After its 2013 launch, Globus Genomics (GG) continued to add new capabilities and users to its cloud-based genomic data service in 2014, helping researchers more easily transfer, share, and analyze sequencing results. This month, GG software engineer and CI fellow Ravi Madduri gave interviews to BioIT World and Amazon Web Services about the software-as-a-service facilitates more rapid and reproducible research.
As a prelude to the World Urban Forum held this year in Colombia, the University of Chicago organized an event called “The Informed City: Data Driven Approaches to a More Just, Equitable and Sustainable City.” The day featured panels highlighting work at the CI’s Urban Center for Computation and Data, the Institute for Molecular Engineering, and the UChicago Crime Lab.
A paper from CI Senior Fellow and Faculty Andrey Rzhetsky computationally reviewed 100 million medical records to conduct an unprecedented nationwide study of the prevalence of autism and related disorders. The results of the study, which came from Rzhetsky’s Conte Center for Computational Neuropsychiatric Genomics, pointed to environmental factors as the most likely cause of the recent increase in autism spectrum disorders over the last two decades.
Our Inside the Discovery Cloud series moved its focus to Particles and Cosmos, with talks from Center for Multiscale Theory post-doc John Grime on modeling HIV and CI Senior Fellow Katrin Heitmann on modeling the expansion of the universe.
Planet & Environment was the theme for this month’s Inside the Discovery Cloud, with talks from Argonne’s Scott Collis on the physics of clouds and Elisabeth Moyer of the Center for Robust Decision-Making on Climate and Energy Policy (RDCEP) on constructing better climate models to inform policy-makers and the public.
How will climate change affect Chicago? That was the central question at a panel organized by Chicago Public Radio, which featured David Weisbach and Moyer from RDCEP. Their answers were bittersweet -- while Lake Michigan provides the city with a precious source of freshwater, “knock-on effects” such as climate refugees from the coasts will challenge Chicago’s infrastructure. Later in the month, Moyer also appeared on WTTW’s Chicago Tonight to talk about the relationship between climate change and national security.
A study by Knowledge Lab found that medical research is disproportionately focused on diseases found in wealthy countries, neglecting the more common diseases that primarily afflict the developing world. The analysis of nearly 4 million scientific articles found that researchers are more likely to study diseases that are more prevalent in their country, leading to a “double jeopardy” for the world’s poor -- the highest health burden from diseases that are studied the least.
It was a good month for Globus, the CI’s research center for developing new, powerful research data management services. First, the Globus data sharing service was adopted by the San Diego Supercomputer Center to help researchers work and collaborate on large datasets. A grant from the Sloan Foundation supported efforts to evolve Globus from free service to sustainable non-profit. And at the annual GlobusWorld meeting at Argonne, CI Director Ian Foster announced new Globus data publication services, which will help scientists permanently publish and share their data, creating more reproducible and accessible research.
Researchers from around the globe gathered at the University of Chicago to discuss the heady topic of what computational tools can do for the science of science, in the first-ever Information, Interaction and Influence workshop. Co-organized by Knowledge Lab and Digital Science, the two-day event covered online profile systems, how to measure scientific impact in the digital age, new scientific software start-ups, and the potential and perils of data-driven science. A full report on the conference was published in August.
The CI and Knowledge Lab launched the Computational Intelligence series, bringing in scholars inventing and applying new techniques to harvesting knowledge from digital information in science and the humanities. The first two talks of the series, from IBM’s Dashun Wang and the University of Maryland’s Yuening Hu, covered methods to predict a paper’s future scientific impact and cooperative human-computer text mining.
Agent-based models are now used by scientists to study everything from epidemiology and traffic to economics and terrorism. A talk by CI Fellow John T. Murphy covered ABM projects underway at Argonne’s Decision and Information Services Division and the development of the Repast modeling software, which allows researchers to scale up agent-based models to supercomputer strength.
The Big Data & Health conference, organized by the CI and the Center for Health and Social Sciences (CHeSS), featured several CI-related speakers, including Senior Fellow Nancy Cox on the computational challenges facing genomic research, Faculty Samuel Volchenboum on an international database for studying the rare disease neuroblastoma, and Senior Fellow Maryellen Giger on the extraction of clinical data from medical images.
For the May edition of Inside the Discovery Cloud, the subject was Life Sciences and the speakers were CI Senior Fellow Rick Stevens, on web-based tools to study and model microbes and microbial communities, and Fellow Gary An, on simulating the human gut to study sepsis and intestinal diseases.
The second edition of the Data Science for Social Good Fellowship began in their downtown Chicago space, bringing in 48 students and 6 mentors to work with non-profit and government organizations on data-driven projects.
A paper from CI fellow Joshua Elliott, RDCEP co-PI Elisabeth Moyer, and graduate student Michael Glotter examined the practice of “downscaling” climate models to connect them with hydrology and agriculture simulations. Their paper, published in PNAS, found flaws in the method, arguing for the pursuit of alternative approaches to multi-sector models of climate change and its impacts.
The new Swift on the block -- Apple’s programming language for iOS devices -- brought renewed attention to the first to bear that name: CI’s Swift parallel scripting language. A feature on Swift talked Senior Fellow and Swift team leader Michael Wilde about the how the high-level language helps researchers in disciplines ranging from theoretical chemistry to urban planning accelerate their big data research.
Protecting people from influenza requires staying one step ahead of the disease as it mutates into new forms every season. CI Senior Fellow Gregory Voth, in PNAS, published a new study that uses multi-scale modeling to study a viral mechanism targeted by the flu drug Amantadine, providing helpful information for winning the race against drug resistance.
The second meeting of the Urban Sciences Research Coordination Network (USRCN), administered by the Urban Center for Computation and Data, provided updates of several projects seeded at the network’s launch in 2013. Researchers presented innovative collaborations using agent-based models to study the flow of health information in urban communities, wrangling cumbersome city data sets into user-friendly portals and maps, and combing through massive student databases to understand the factors that lead to academic success or failure.
Globus joined a multi-institutional effort to develop a “data facility” for materials science, a central repository where researchers can store, publish, and share experimental data, driving new collaborations and advancing the discovery of new materials. The project is a pilot for a broader National Data Service, a new infrastructure to bridge the data silos at universities and laboratories across the United States.