Campus Champions Fellows Projects 2013
Check box next to name to compare projects. Reload page to see full list again.
- Abstract:Lack of proper thrombus formation can result in bleeding disorders. Conversely, excessive pathogenic thrombosis and/or embolism can lead to temporary or permanent obstruction to blood flow, which results in over 1 million heart attacks or strokes each year in the US. Recently, in-vivo studies have reported that the thrombus structure is heterogeneous, consisting of a compact core of activated platelets near the blood vessel injury and a loose shell of quiescent platelets on the outer periphery of the thrombus. Motivated by the recent ability of intravital microscopy experiments to image the thrombus structure with a single platelet resolution, we set out to study the microenvironment within the heterogeneous thrombus structure, as well as to try and identify any possible implications that it could have on drug target design. The approach implemented in this study is to: perform in-vivo laser injury experiments in mouse cremaster muscle arterioles, image the resulting thrombus structures using fluorescent confocal microscopy with 0.216 micron resolution and model the fluid dynamic environment experienced by the thrombus using supercomputer Lattice Boltzmann method (LBM). Additionally, tracer tracking simulations are performed in order explore diffusive transport of different-sized coagulation factors within the thrombus, released either from the injury site or from the lumen. Although the exact mechanism is unknown, it is hypothesized that once platelets are recruited from the blood stream to the injury site, a complex interplay between their activation state (degranulation/shape change), transport of coagulation factors and forces due to blood flow is responsible for determining the thrombus structure and integrity. Thus, the goal of the proposal is to study how transport of coagulation factors (Specific Aim I) and forces exerted by the blood flow (Specific Aim II) relate to the heterogeneous thrombus architecture. The proposed innovative multidisciplinary approach is expected to yield an enhanced fundamental understanding of thrombus structure and how mechano- and chemo-perturbations affect it. Collectively, the outcomes of this work will provide insight towards understanding clot generation, structure and stability relevant to thrombosis, thromboembolism and coagulopathy, ideally to reduce risks of occlusion and embolism.
- Abstract:In this research proposal, we request an allocation of service units on various platforms to perform calculations, using state-of-the-art computing algorithms for the study of nanostructured materials. Specifically, apart from the in-house developed first-principles molecular dynamics code and density-functional tight-binding molecular dynamics algorithms, we plan to employ the commercial code VASP (Vienna Ab Initio Package) and GPU-based quantum chemistry code. The computational study will be crucial for supporting the ongoing research project in multiscale modeling of nanomaterials.
- Abstract:A modified version of the ROMIO ADIO Lustre driver will be tested for performance and correctness on large-scale systems using widely available parallel I/O benchmark programs (for example, FLASH I/O and BTIO). The code to be tested includes an implementation of an efficient file domain partitioning algorithm for Lustre file systems, as well as the replacement of several blocking MPI calls with non-blocking variants. The implementation of file domain partitioning is based on the "group-cyclic" algorithm presented in a paper by Liao and Choudhary, "Dynamically Adapting File Domain Partitioning Methods for Collective I/O Based on Underlying Parallel File System Locking Protocols". The modified code has previously been tested in a small-scale setting over Infiniband and 1 Gbit Ethernet, where it has been validated, and has been shown to increase write throughput by up to a factor of three over the original ADIO Lustre driver code. Additional code modifications based on one-sided MPI communication will also be tested. Ultimately, this testing will be used to support the inclusion of the code changes into the ADIO codebase.
- Abstract:The key features of 2010 U.S. health insurance reform are the government 's substantial involvement on the private health insurance markets: it creates a government regulated subsidized health insurance market and imposes tax penalties to the uninsured individuals and large employers not offering health insurance. This project presents a structural estimate of those impacts on individual health insurance coverage, health care spending, and health and labor market outcomes, investigates how these impacts differ across different individuals, and explores whether there exists a welfare enhancing alternative reform. We develop and empirically implement an equilibrium labor market model where individuals make health care and labor supply decisions over the life-cycle and employers make health insurance coverage decisions. The model allows rich heterogeneity both worker and employer sides, and is estimated by using various micro data sources through simulated method of moments. The numerical procedure is two-fold. First is calculation of the objective function for a given vector of parameters. Second step is minimization of the objective function by searching for various parameter values. Our numerical algorithm is written in FORTRAN 90, and we implement the estimation by using derivative-free parallel optimization procedures, such as HOPSPACK.
- Abstract:There is a national need to provide researchers, educators and resource managers with seamless and fast access to essential geospatial/geotemporal data, physics-based numerical models, and data fusion tools that are necessary to predict and manage the nations surface and groundwater resources. The evaluation of ecosystem and watershed services such as the detection and attribution of the impact of climatic change provides one of many examples of the pressing need for high resolution, spatially explicit resource assessments. However, the current cyber infrastructure for supporting models and data at a national level in the United States must overcome two important problems: 1) Accessibility: At present there is no frameworks that allows simultaneous access to the necessary regional data and high resolution numerical watershed modeling tools. The essential data resource itself (climate reanalysis products, stream flow, groundwater, soils, land cover, satellite data products, etc.) resides on many servers such that fast and efficient access to the data during model development, analysis and simulation is not yet feasible, and 2) Scalability: Given a cyber infrastructure for a model and data integration, the problem of scalability emerges as the next crucial issue. How can we align and support evolving regional geospatial and geotemporal data products with new watershed model parameterizations, support regional versioning of models and new data sources while also offering the benefits of these new services to scientists, resource managers and stakeholders with regional, national or global interests?
- Abstract:Systematic differences or biases that exist between observed and modeled climate obscure quanti- tative measures of scientific credibility in model predictions. In fact, without well-justified observational tests of the key unknowns, the question of how uncertainty should be defined is ill-posed. One manifestation of this conundrum is that the climate change community has not been able to come up with a metric of model skill that can anticipate how different climate models will respond to greenhouse gas forcing. This ambiguity also impacts model development as it can be unclear which model biases are affecting predictions from those that do not. The point of the proposed calculations is to advance the computational and statistical framework that can be used to investigate how observational data should be used to improve representations of climate processes, quantify model credibility, and define uncertainty.
- Abstract:We propose to develop and demonstrate XSEDE enabled national scale ABM (Agent Based Modeling) simulations of the spread of infectious disease. This work is part of the NIGMS sponsored MIDAS program that has been tasked with developing national scale modeling tools for pandemic preparedness. And important issue in these types of models is their scalability in terms of large populations. Previous work using the Imperial Model reported by Ferguson et. al. has shown that large scale simulations are possible. However, the general unavailability of large SMP machines has limited the time to complete an experiment involving many hundreds of runs. We foresee that the SMP computational resources located at the PSC would allow us to accelerate the generation of results using synthetic populations representative of the population of the United States. This study therefore offers the opportunity to explore both the usefulness of large scale simulations and the appropriateness of XSEDE resources for models of very large size.
- Abstract:Considering we have several mobile networks data on hand, we would like to do some comparative study on them. The goal is to have a paper on discovery. In the first step, we would like to conduct a set of experiments on the following metrics: Here are a few of network statistics: 1. degree distribution 2. diameter of graph 3. total calling time 4. weight distribution on weighted graph (w/ call duration as weights) 5. eigen values 6. diameter over time 7. number of triangles 8. number of nodes in core (largest connected components) 9. max degree In the second step, we would like to do some network behavior analysis: 1. Fitting: lognormal, loglogistics, DPLN, etc. 2. Generation process: CRP and several of my models (e.g. RCRP) 3. New-comer, churner and consistent users' behaviors analysis In the third step, we would like to do the interpretation for the similarity and difference from the analysis of Step 1 and Step 2.
- Abstract:Our goal is to develop efficient algorithms to combine NASDAQ order submission, cancelation and execution messages to generate message-by-message snapshots of the financial market with nanosecond resolution. The construction of this data makes it possible to analyze the behavior of computer-based high frequency trading and suggest possible regulatory solutions. Efficient algorithms run on high-end compute resources enable researcher and policy markets to rapidly reconstruct trading activity and analyze suspicious behavior or unusual market events. This effort will be the largest computing project to date in academic finance and the output of the project will be about 200-300 terabytes in size. Our research using Blacklight and Gordon have already had policy impact. On September 7, Bloomberg news mentioned that the U.S. regulators decided to vote on the policy recommendation in one of our papers using supercomputer. On September 20, the U.S. senate discussed another paper of ours using supercomputers at Capitol Hill , and in the testimony, it is quoted as In a ground-breaking study done at the University of Illinois at Urbana-Champaign, Mao Ye, Chen Yao and Jiading Gai demonstrated the following: Our paper has already between accepted both by the primary conference in finance (American Finance Association Annual Meeting) and supercomputing (XSEDE 2012).
- Abstract:Protein translation is one of the most fundamental and universally conserved biological processes. As a result a detailed understanding of the steps involved remains a major research focus in molecular biology and cellular biology. Most of our knowledge of protein production is limited to results from laboratory experiments. However, if selection on translational errors and efficiency contribute to the evolution of the codons used to create a protein, it follows that genomic data contains information on the nature and importance of these forces. A major challenge facing the scientific community is how to extract this information in an efficient and biologically meaningful manner. The goal of this research project is to develop statistical techniques that will allow scientists to efficiently extract and interpret this biologically important information from an organisms genome. The major outcomes of this work include the ability to estimate key biological parameters such as the direct cost of protein translation, the frequency at which different errors occur, and the average expression level of every gene within a genome.
- Abstract:Rotation, stratification and turbulence are all present with different amplitudes at different scales in the atmosphere and in the oceans of Earth, with a set of dimensionless parameters that is reachable numerically as far as inertial and gravity waves are concerned, although the Reynolds number of geophysical flows is still too low in direct numerical simulations (DNS) unless one uses adequate parametrizations of small scale interactions. Breaking of gravity waves is important, together with Kelvin-Helmoltz (KH) roll-up, for example as a source of clear-air turbulence in the atmosphere, or of enhanced mixing as in the abyssal Southern Ocean at mid latitudes. with, In the oceans in general, sub-mesoscale frontogenesis and significant departure from quasi-geostrophy are seen as the turbulence intensifies. Few recent studies, if any, have considered specifically the detailed role of helicity (velocity--vorticity correlations) in geophysical flows, although it was recognized early on that it may be important in the atmosphere as well as in the generation of magnetic fields. Helicity has been measured recently in the atmosphere and its role has been evaluated for example in hurricanes. The questions we shall be dealing with in this proposal are: How and where is helicity created in such flows? What is the role of helicity on structures, nonlinear dynamics and statistics? What is the interplay between geostrophic balance, potential vorticity conservation and the creation of helicity? Does the production of helicity alter the mixing properties of such flows? How can the role of helicity be modeled efficiently? To help address these and many other turbulence questions, we have developed a code (Geophysical High Order Suite for Turbulence, GHOST) which can perform DNS and several types of Large-Eddy Simulations (LES). The code parallelizes using a hybrid method for grids up to 6144^3 points on up to 98000 compute cores, and, together with simulation data, is available to the community. In order to study the problems mentioned here, we request 37 million core-hours to perform DNS of forced helical stratified rotating turbulence on tri-periodic grids of up to 3072^3 points to a steady state regime, as well as to conduct runs at lower resolutions continuing our exploration of parameter space, using both DNS, and a LES model, which has been validated previously for rotating helical turbulence on grids of up to 3072^3 points. Part of this request will also test the scalability of the hybrid scheme up to ~100,000+ cores, as well as to characterize the LES at large core count. Broader impact and educational component: Rotation and stratification are both present in most geophysical and astrophysical flows. A better understanding of their effect on small-scale turbulence, and vice-versa, is an essential piece of weather and climate modeling, as they affect cloud formation or vertical mixing and its influence on global meridional overturning circulation; they are also an essential piece of the puzzle of magnetic field generation in planets, stars and galaxies. The code, in which a variety of other solvers (e.g., Surface Quasi Geostrophic, Shallow Water, or magnetohydrodynamics) are provided, and the data will all be made available to the community. A particular effort will be made, as with our preceding allocations, to immerse and graduate students and post-doctoral fellows in the numerical and modeling techniques for geophysical and astrophysical multi-scale turbulence by way of computation, analysis and visualization of the runs described herein, at the frontier of high performance computing.
- Abstract:This is a request for continued access to XSEDE resources to support the NSF and NIH funded research efforts in the Cheatham lab at the University of Utah. This work involves breaking and fixing atomistic biomolecular force fields and molecular dynamics methods with a focus on proteins and nucleic acids. Our goal is to better understand biomolecular structure, dynamics, interactions and free energetics and ideally conformational changes due to subtle shifts in the environmental conditions (solvent, salt, ligands).
- Abstract:This proposal requests computing time on Stampede and Gordon for the purpose of performing atomistic simulations (1) to compute surface energies of selected stable and metastable phases of alumina over a range of temperatures and (2) to study damage induced by ion bombardment on a subset of the alumina surfaces. Molecular dynamics simulations will be performed on Stampede using LAMMPS with the ReaxFF potential for Al-O systems, while virtual X-Ray diffraction (XRD) and selected area electron diffraction (SAED) patterns of select atomic configurations will be performed on Gordon, using a new computational procedure recently proposed by the PIs. In total, approximately 1.7M SUs are requested on Stampede and 275,000 SUs are requested on Gordon to perform the research presented in this proposal. This research is supported by the National Science Foundation under grant CMMI#0954505.
- Abstract:This allocation will be used to perform continuum-level simulations of microstructures in materials (currently funded by the National Science Foundation (NSF) under DMR0746424 (CAREER), DMS0854905 (battery simulations), DMR0907030 (fuel cell simulations) and DMR1105409 (phase field crystal model development), as well as three Department of Energy (DOE) grants). The proposed work consists of seven projects: (1) Simulations of solid oxide fuel cell electrodes, (2) Simulations of lithium ion intercalation compounds, (3) Phase-field-crystal simulations of nanostructures, (4) Simulations of gallium nitride growth by selective area epitaxy, (5) Simulations of anodic Al2O3 nanopore growth, (6) Three-dimensional characterizations of complex microstructures, (7) Coarsening of complex three-dimensional microstructures. We develop and/or adopt state-of-the-art models and numerical methods, and therefore the majority of the codes have been written by the participants using FORTRAN with the Message Passing Interface (MPI) Library. With the past and current support, we have generated two Ph.D. theses, four published or accepted journal publications (among them is a Nature Communications paper), three submitted manuscripts, and eight manuscripts in preparation.
- Abstract:In response to the emerging research needs identified by National Research Council and Intergovernmental Panel on Climate Change, with support from NSF, we are developing an Integrated Technology-Driven Earth System Model (ITDEaSM) to link technology options, energy use, and policy choices and resultant emissions with Earth system processes that govern climate change, air quality, water quality, and ecosystem. ITDEaSM will be developed by coupling and adding capability to a suite of community models to address grand challenges in decadal regional climate predictions and fill in major gaps in existing climate, Earth system, and ecosystem modeling. Multi-decadal simulations will be performed using ITDEaSM to estimate the long-term impacts of global climate changes on the global-through-urban earth system, including air quality, water quality/supply, forest, and ecosystem. Our ultimate goal is to quantify such impacts, reduce associated uncertainties, and identify technology choices for co-benefits of climate/Earth system mitigation. During this 1-year allocation time period, our work will focus on model improvement, testing, and initial applications.
- Abstract:There are four major projects for which we are requesting continued used of XSEDE supercomputer resources. The first project (P-1) is to examine the manner in which ions affect the conformational stability of proteins. Both experimental and computational work has shown that ions influence the conformation of protein in aqueous solutions, yet the physical mechanism behind this phenomenon is not well understood. The proposed work will examine how both water molecules and ions interact with small peptides. The goals of this project are to utilize computational methods to study interactions between ions, waters, and various moieties of small peptides to elucidate the physical processes responsible for altering peptide conformations in aqueous solution and to identify the most stable conformations of these peptides in various concentrations of biologically relevant salts. The second project (P-2) will involve determining the free energy landscape of polyQ which will be compared against the experimental conformational free energy landscape. The third project (P-3) will investigate the binding free energy of substrates and inhibitors of monoamine transporters (MATs). Finally project P-4 will continue our work into the dynamics of MAT proteins. For the last project we are requesting ECS support to implement a new boundary condition into Amber in order to more efficiently study membrane bound proteins by simulation.
- Abstract:Proteins play a key role as therapeutics in a number of diseases and protein crystallization is a central activity in the pharmaceutical industry. Specifically, the production of highly-ordered, high-quality protein crystals through batch crystallization processes is vital in devising proteins for therapeutic purposes. However, despite extensive experimental and computational work on understanding protein structure and function, there is a lack of a systematic framework that relies on a fundamental understanding of the nucleation and growth mechanisms of protein crystals at the microscopic level and utilizes such information to model and operate protein batch crystallization processes at the macroscopic level. Motivated by the above considerations, a hierarchical and computationally tractable approach will be implemented to: (a) elucidate the phase diagrams and understand the physics of crystallization of globular proteins at the microscopic level, (b) deduce microscopically consistent crystal nucleation and growth rate laws, and (c) utilize the previous knowledge to model and control batch protein crystallization processes at the macroscopic level. The previous objectives will be achieved by the following specific projects: (1) The equilibrium protein phase diagrams will be determined via coarse-graining techniques, Monte Carlo simulations, and finite-size scaling theory. The simulations will be implemented according to the methodology of spatial updating, recently proposed by the PI. Crystal nucleation will be studied by a combination of Monte Carlo and molecular dynamics simulations. (2) The results of part (1) will be used to: (i) study the growth of protein crystals via molecular dynamics and kinetic Monte Carlo simulations, and (ii) derive microscopically consistent rate laws. (3) The nucleation and growth rates of parts (1) and (2) will be used to model batch protein crystallizers through population and mass and energy balances. These models will be used in conjunction with stochastic model predictive control methodologies that account for model uncertainty to achieve a desired crystal size distribution at the end of the batch protein crystallization process. The theoretical results will be tested and validated against available experimental data. The present work will provide a basic understanding of the factors that affect protein phase separation and crystallization and will benefit a wide range of pharmaceutical companies and products. In addition, it will provide a practical framework to simulate, model, and control globular protein crystallization processes. The research program is well-suited for training graduate as well as undergraduate students in the area of molecular and multiscale modeling and simulation.
- Abstract:The most important and complex decisions in governments, businesses and academia are made in teams. And yet, as recent events such as the hurricane Isaac in Louisiana, Benghazi attack in Libya, and the global financial crisis indicate, assembling effective teams at short notice is a daunting task, for intellectual as well as logistic reasons. We still have sparse socio-technical knowledge of how potentially globally distributed cross-cultural teams and systems of teams are assembled, or how a given mode of assembly impacts effectiveness. This project is to address this limitation. We attempt to develop a methodology to understand the socio-technical dynamics shaping the assembly of teams and ecosystem of team in distributed global contexts. While there is a vibrant body of related research focused on developing theories and tools to enable and understand how teams collaborate, the current effort is more specifically focused on understanding a precursor to the collaboration itself - the mechanisms by which collaborations are assembled. Due to the computation complexity, doing this type of research will require high performance computing resources. An important barrier to understanding team assembly is finding a suitable research environment. This environment would be one in which geographically distributed individuals from potentially different cultures are assembling in teams of varying sizes to accomplish a variety of tasks over varying durations. It would be one in which their actions, interactions and transactions are captured with precise time-stamps. Finally, their individual, team, and team ecosystem outcomes would be recorded with well-defined metrics. EVE Online, a massively multiplayer online role-playing game, offers a suitable research opportunity to study the assembly of teams and ecosystems of team. This project will employ three different analytical approaches to test models of team assembly, p* network analysis, SIENA network modeling, and hypergraphs. Analysis of thousands of nodes is computationally demanding, and XSEDE resources and ECSS support are required. The project will advance theory in the areas of group dynamics, contribute to our understanding of analysis of large scale networks, advance computational methods for network analysis, and enable the comparison of different methodologies.
- Abstract:The goal of this effort is to arrive at a highly accurate assembly of the Timber Rattlesnake (Crotalus horridus) genome by employing the latest advances in technology with respect to sequencers, assembly software, and high performance computing. The potential for valuable medical breakthroughs is substantial. Also, a complete genome assembly of the species will lead the way to less need for others to attempt de novo assemblies as reference genomes will be available. An additional benefit of this work will be to improve processes and methods for others that will soon be attempting whole genome assemblies. As XSEDE ECSS support has been requested along with computational resources, it is expected that the ECSS staff will become a conduit for the lessons learned, thus disseminating this information most efficiently to others in the specific field of research.
- Abstract:The largest source of uncertainty in response to a temperature forcing among climate models is the shortwave cloud feedback. Our ability to model this feedback is greatly hindered by the inherent limitations of our climate models to simulate the sub-grid scale convection responsible for generating low clouds. As a result, low clouds must be parameterized, which results in a large discrepancy in feedback between climate models. This discrepancy in turn leads to a wide variety of possible future climates. Using XSEDE resources, we intend to improve our ability to project climate change by refining shortwave cloud feedback using observations coupled with modeling. We will combine remotely sensed cloud data from a variety of space-borne platforms to create a model of the atmosphere and its constituents. We will then use plane-parallel radiative transfer calculations to derive top of the atmosphere radiation using the Rapid Radiative Transfer Model (RRTM) and compare to observations. Once the model has replicated the observed top of the atmosphere radiation, we will apply perturbations to cloud properties to determine the sensitivity of the shortwave feedback to perturbations in cloud properties. This approach, while expensive in terms of radiative transfer calculations, removes the effects of climate model adjustment bias and allows us to restrict the range of possible shortwave feedbacks on the climate system due to cloud changes based on observed quantities. The insight from measuring the radiative forcing due to various cloud property changes will greatly aide not only in restricting current climate modeled cloud feedbacks, but will also yield insight into the development of new cloud parameterization algorithms in climate models.
- Abstract:We are exploiting HPC resources to enable high-throughput image understanding in dental computing.
- Abstract:The University of Alabama in Huntsville developed Algorithm Development and Mining Toolkit (ADaM) has been successfully used to solve various science problems. This project will experiment with the use of XSEDE resources to execute data mining workflows on big data. We propose to utilize the Apache Airavata software for workflow composition, monitoring, and execution. We expect the project will help scientists who do not have access to computing resources to seamlessly create, execute and share data mining workflows.
- Abstract:Heliospheric space weather represents changing conditions of plasma and magnetic field in the interplanetary space that can influence spacecraft systems, communication and navigation, power grids, and humans. Because of the sparse measurements and the vast spatial dimensions, numerical modeling plays a critical role in clarifying and predicting impacts of solar wind and solar eruptive phenomena. Accurate computation of the solar wind parameters is crucial for predicting co-rotating stream structures and for transient disturbances (interplanetary shocks and coronal mass ejections) that propagate and interact with those background structures. We have developed the 3D numerical magnetohydrodynamic code ENLIL which uses coronal maps and fitted coronal mass ejections to predict corotating and transient disturbances in the inner heliosphere. The code is written in Fortran 90, uses MPI for domain decomposition, and NetCDF for output data files. We are requesting a Startup Account to further develop ENLIL code. Current solar observations are not sufficient to drive numerical simulations and various model assumptions and model free parameters are used. Large-scale model calibration and validation is needed to determine optimal parameters that would provide reasonable prediction accuracy within the whole solar activity cycle (11 years). This requires computational resources beyond the multi-processor workstation and small cluster for code development and forecasting application.
- Abstract:Understanding regional scale water resource systems requires understanding coupled hydrologic and climate interactions. The traditional approach in the hydrologic sciences has been to either treat the atmosphere as a forcing condition on the hydrologic model, or to embed a hydrologic model in a climate model. We have been pursuing a different approach that couples hydrologic and climate models through web services. Last year our team successfully prototyped two-way coupling using this technique; this year we wish to advance the work by validating the models scientifically.
- Abstract:Spatiotemporal thinking and analysis has been a common interest in a growing research community aiming at understanding the spatial patterns of socioeconomic trends and the dynamics in the changing geographical structures. While prior theories on socioeconomics and urban growth cover the space-time dimensions and there is an increasing awareness of its importance in the empirical analysis, the rich details of space-time complexity remain largely unexplored because of the constraint in computation capacity in response to the challenge of data intensive computing. As a result, the current space-time simulation and statistics in such research can only deal with a limited amount of data. We would like to explore high performance solutions targeting two data and computational intensive problems in criminal analytics and urban growth simulation.
- Abstract:This proposal is a renewal request for resource allocation from XSEDE to support our ongoing research in computational fluidstructure interaction (FSI) in two areas of application. FSI is ubiquitous in biological and biomedical systems. In our work, we focus on two different applications by using the same computational methodology. The first one is to investigate the aerodynamics of flexible flapping wings and the flight dynamics of flyers in nature. The study will help develop highly agile biomimetic micro air vehicles that may fly like insects, birds, or bats. The second one, a new subject added to our current allocation, is modeling of the biomechanical interaction between the glottal airflow with the vocal fold, and the study will help develop a future computational tool that can be applied in the clinical management of voice disorders. For this renewal proposal we will pursue three specific projects that fall within the scope of those two research themes: (1) direct numerical simulations of hummingbird flight; (2) high-fidelity modeling of aeroelasticity in the cicadas forewing; and (3) the role of nonlinear tissue mechanics in the flow-induced vocal fold vibration. The proposed numerical study will significantly advance computational modeling of FSI in biological and biomedical systems.