2017-2018 Blue Waters Graduate Fellows

Ten outstanding computational science PhD students from across the country have been selected to receive Blue Waters Graduate Fellowships for 2017-2018. The fellowship program, now in its fourth year, provides substantial support and the opportunity to leverage the petascale power of National Center for Supercomputing Applications (NCSA) at the University of Illinois' Blue Waters supercomputer to advance their research. The awards are made to outstanding PhD graduate students who have decided to incorporate high performance computing and data analysis into their research.

Matthew Clement
Matthew Clement
University of Oklahoma

The Solar System is a fascinating dynamic system with planets' orbits changing chaotically over millions of years. To date, computer simulations have been failing at reproducing formation of Mercury and Mars of the sizes and masses they are today. Using iterative techniques, Matthew Clement will probe whether an early dynamical instability between the giant planets could scatter enough material away from a forming Mars to produce a planet of the correct mass. Matthew will use Blue Waters to run thousands of computer simulations using the state of the art code to uncover the past evolution and verify the long-term stability of the Solar System with an emphasis on the dynamics of the four terrestrial planets: Mercury, Venus, Earth, and Mars.

Salme Cook
Salme Cook
University of New Hampshire

Cohesive sediment containing high levels of pollutants that are responsible for decreased water clarity and quality are found on coastlines throughout the World. Environmental managers and scientists that guide legislative policies and conservation efforts are increasingly relying on computer simulations that take into account the hydrodynamics, sediment transport, and associated biogeochemical fluxes. Limitations of these models typically include accessibility of computational resources to resolve these processes at the necessary temporal and spatial scales, and availability of observational data to verify model results. Salme Cook from the University of New Hampshire will use Blue Waters to understand the importance of wind-wave induced sediment transport in driving nutrient fluxes in cohesive sediment environments using a coupled hydrodynamic-sediment transport model.

Evan Feinberg
Evan Feinberg
Stanford University

Pathogenesis of Alzheimer's disease (AD) and associated chronic pains continue to draw the attention of scientists and pharmaceutical companies. The beta 1 adrenergic receptor is a novel target for AD therapies and the μ-opioid receptor is the primary target for clinically used opioid analgesics. Both proteins are G protein-coupled receptors (GPCRs) that sample many functionally important conformational states. In this project, Evan Feinberg will use Blue Waters to conduct molecular dynamics simulations of these biomedically critical proteins. Using Blue Waters' GPU-enabled nodes, Evan will improve a deep learning algorithm for predicting protein-ligand binding affinities. Achieving these goals will illuminate the structural biology of GPCRs and enrich the search for new therapeutics for AD and chronic pain.

Lauren Foster
Lauren Foster
Colorado School of Mines

Characterization of climate feedbacks in snowmelt-dominated headwater systems is critical to predicting water supply for more than one-sixth of the World's population. Management decisions for water yield in downstream basins are, however, largely informed by parametrized, low-resolution models that do not resolve the non-linear feedbacks driving behavior in topographically complex regions. Using the Blue Waters system, Lauren Foster will apply a hyper-resolution (10 m) integrated model over 255 square kilometers to perform the first study to vary resolution over 2 orders of magnitude. This will provide an opportunity to bracket uncertainty in projected change and to determine the scale at which functional hydrologic relationships break down.

Photograph image of Zachary Goldsmith
Zachary Goldsmith
University of Illinois at Urbana-Champaign

NiFe oxyhydroxide is a highly active electrocatalyst for the oxygen evolution reaction, a crucial process for carbon-neutral energy storage. Quantum chemical calculations have confirmed the operando Mössbauer spectroscopic identification of Fe(IV) sites in the material. However, the nature of the catalytically active site—the metal site, its oxidation state, and its coordination—has yet to be determined with the same rigor. Zachary Goldsmith will use Blue Waters to perform electronic structure calculations of Ni(Fe) oxyhydroxide edges in concert with in situ spectroscopy to elucidate the chemical and electronic environments most conducive to facilitating catalysis. This work will push the boundaries of modeling reactive interfaces and establish novel catalyst design principles.

Jennifer Hays
Jennifer Hays
University of Virginia

Flexible recognition is a common paradigm in infection; many pathogens have proteins that are structurally and mutationally flexible but still bind human cells. Determining the structural basis of this recognition is computationally difficult as the receptor-bound state can comprise many different structures. Current methods that incorporate experimental data into molecular dynamics simulations can refine structures of rigid proteins but cannot capture large backbone changes of dynamic proteins, especially for complex experimental data. Jennifer Hays will apply a newly developed methodology that leverages petascale computing resources provided by the Blue Waters system by simulating many ensemble members restrained by complex spectroscopic data to refine conformational ensembles of flexible proteins.

Ethan Kruse
Ethan Kruse
University of Washington

When NASA's Kepler telescope failed, it was revived as the K2 mission but with degraded data quality and no working pipeline to search for exoplanets. Every 80 days K2 observation "campaigns" produce new data for different regions of the sky and, hence, with different sets of stars. Ethan Kruse, co-developer of the best data reduction and exoplanet discovery pipeline for K2, will use Blue Waters' facilities to more efficiently process and release planets' catalogs for K2 campaigns as they are published. Additionally, if time and resources permit, Ethan will search the original Kepler data for a rare but important class of planets the official pipeline may have missed (those with transit timing variations).

Rachel Kurchin
Rachel Kurchin
Massachusetts Institute of Technology

Photovoltaic (PV) device performance is frequently limited by the presence of bulk point defects. Rachel Kurchin from the Massachusetts Institute of Technology will use Blue Waters to understand the structure and physics of these defects through a series of density functional theory calculations. She then plans to use the obtained knowledge to develop generalizable criteria for identifying and/or designing the so-called defect-tolerant materials. Such materials are expected to minimize defects' detrimental impact on device performance. Kurchin's work will open the doors to the use of lower-cost manufacturing techniques which typically tend to introduce more defects. This, in turn, will result in the faster scale-up and broader deployment of PV required to combat the massive challenge posed by the climate change.

Rachael Mansbach
Rachael Mansbach
University of Illinois at Urbana-Champaign

Self-assembling π-conjugated peptides are attractive candidates for the fabrication of bioelectronic materials possessing optical properties due to electron delocalization over the conjugated peptide groups. A prerequisite to the rational design of biomolecules for such applications is a molecular-level understanding of aggregation. In this work, Rachel Mansbach will employ Blue Waters to perform a multiscale computational investigation to understand and design a family of self-assembling peptides with tunable optoelectronic responses. Such an investigation requires intensive computational resources due to the need to characterize and engineer properties at the quantum, molecular, and supramolecular levels.

William Payne
William Payne
University of Nebraska Medical Center

Developing computational tools to aid in the rational design of polymeric drug delivery systems could dramatically improve the therapeutic outcome of many diseases, especially cancer. The enhanced pharmacokinetic profile obtained through nano-formulation provides higher accumulation in target sites and decreased comorbidities. William Payne from the University of Nebraska Medical Center will use Blue Waters to develop a library of computational models for amphiphilic polysaccharide nanoparticles. This library will enable formulation optimization for a wide variety of drugs. Upon validating these models experimentally, the developed toolset will be used to optimize the formulation of an anticancer chemotherapeutic.