Chi-Ren Shyu - US grants

The University of Missouri-Columbia is awarded an early career development grant for Dr. Chi-Ren Shyu to develop new approaches to knowledge federation of visual phenotypes that can function and flourish in today's landscape of diverse, constantly changing and inconsistent databases, schema, and terminology. The knowledge-sharing hub proposed in the project has the ability to use disparate computational resources to reliably exchange data and computations with little intervention from the user. A second challenge is to develop methods that answer the researchers' hybrid queries, which are combinations of scientific expertise and algorithms that detect and retrieve information related by image features, semantic description, map location, and patterns of DNA sequences, even when those data are not currently correlated with each other in the partner's databases. The proposed hybrid query system will be the first and only in the research community that allows a researcher or an educator to submit an image of a mutant to the database and ask, "What genes or environmental factors cause this visual phenotype?" The [roject includes an educational component with two unique aspects: One focus is on training of undergraduate and graduate students and creating courses for them that are closely aligned with the research activities. The new courses, titled "Content Management in Bioinformatics," and "Computer Vision and Image Understanding in Bioinformatics" will be pioneering courses in computational biology curriculum design. The other focus will leverage the proposed research activities to create, in conjunction with traditional computer science courseware, scientific data repositories for computation and biology education.

In fields where visually based tacit knowledge is central to decision-making, such as image analysis within the medical or intelligence communities, the difficulty of transferring hard-won expertise to novices is compounded by a lack of sufficient knowledge transfer tools. Static media and current computational approaches are not up to the task of capturing highly complex visually based tacit knowledge. A visual knowledge infrastructure based on enhanced computational approaches and robust tools must be developed that can harness the extant wealth of visual-knowledge expertise
and will facilitate the rapid and effective transfer of that knowledge to and use by learners. Five necessary computer science research goals must be reached in order to build such a visual knowledge infrastructure. They include the development of a declarative knowledge module for visual and textual content management, visual mining algorithms for discovering associations in a multi-dimensional feature space, a cognitive model for visual media analysts, a case-based reasoning system for knowledge repositories and acquisition, and a digital library with a robust knowledge exchange framework for interdisciplinary communities. The cognitive model of visual media analysts proposed in the project has the ability to index facts and production rules from associations among low-level visual features and high-level semantics. To make the machine-readable visual knowledge accessible and sharable by novices, the proposed case-based reasoning system will provide the functions for knowledge retrieval. An efficient visual-knowledge infrastructure is necessary to meet the needs of high quality and cost-effective basic science research and health care. Other impacts include joint mentoring of interdisciplinary students, and the creation of digital libraries for curriculum design in multidisciplinary training, as well as the engagement of underrepresented high school students and undergraduates in the research.

Collaborative Research: Biological Shape Spaces, Transforming Shape into Knowledge

This project will develop a framework to represent, analyze and interpret shapes extracted from images, supporting a wide range of biological investigations. The primary objectives are: (1) to develop a mathematical framework and computational tools for the quantification and analysis of shapes; (2) to integrate these computational models with machine learning and statistical inference methods to enable new discoveries, transforming imaging data into biological knowledge; (3) to deliver novel quantitative methodologies for shape analysis that start from a biological premise, rather than a purely geometric one. The aim is thus not only to quantitatively describe shape, but to develop methods for linking morphological variation to its underlying biological causes. To ensure that the project focuses on methods that are most promising to biology with significant breadth of application, model and tool development will be guided and supported by a set of diverse case studies, ranging from the sub-cellular to organismal scales.

Shape represents a complex and rich source of biological information that is fundamentally linked to underlying mechanisms and function. However, shape is still often examined on a qualitative basis in many disciplines in biology, an approach that is time consuming and prone to human subjectivity. While ad hoc quantitative methods do exist, they are often inaccessible to non-experts and do not easily generalize to a wide variety of problems. The inability of biologists to systematically link shape to genetics, development, environment, function and evolution often precludes advances in biological research spanning diverse spatial and temporal scales, from the movement of molecules within a cell to adaptive changes in organismal morphology. The primary goal of this project is to develop a new suite of widely applicable quantitative methods and tools into the study of biological shape to address the significant need for consistent and repeatable analysis of shape data.

DESCRIPTION (provided by applicant): Researchers and clinicians seek to improve outcomes for patients with and at risk for lymphedema (LE), a chronic, progressive disease of the lymphatic system, often starting with seemingly innocuous superficial swelling of a limb which progresses to a distressing, disabling condition. LE can be life-threatening when infection in LE-affected areas becomes systemic. The toll that inadequately-managed lymphedema takes on individuals'quality of life and the financial impact on the health care delivery system is considerable. Therefore, to evaluate appropriate health care services for patients with all forms of lymphedema and advance the quality of lymphedema care, the American Lymphedema Framework Project (ALFP) launched a national collaborative initiative under the leadership of recognized clinical experts and investigators in the field of lymphedema. However, there is no cyber framework to share findings and knowledge between researchers and practitioners, or to bring the latest findings to patients and other stakeholders. To tackle this challenging issue, our objective for this proposal is to develop cyber informatics tools for the lymphedema stakeholders. The overall goal of this project is to provide an operational cyber infrastructure that collects, organizes, and disseminates up-to-date LE information. This development proposal has three major aims: Aim 1: Develop a cyber framework for inter-institutional lymphedema research activities;Aim 2: Integrate informatics tools to provide an on-line summary of concurrent LE studies;and Aim 3: Create a web portal for lymphedema stakeholders with user-specific query methods Our development plan includes eight development modules for this project: 1) Building a data warehouse for a minimum data set and data governance protocols for cross-institutional studies, 2) Developing survey and analysis tools for each stakeholder group, 3) Computationally collecting high-quality and evidence-based knowledge from a selected set of LE organizations, journals, and news releases from reputable public media, 4) Constructing an informatics tools library for mining structured and unstructured information sources, performing statistical tests, and creating graphics/figures for information visualization, 5) Building a case-based library for complex case repository and exchange within and across stakeholder groups, 6) Building a knowledge base for indexing discovered patterns from all information sources and linking it with the best practice guidelines, 7) Developing user-centered query methods dedicated for patients/families, health professionals, health educators, and researchers, 8) Monitoring the proposed work by quantitatively and qualitatively measuring the improvement of research/clinical outcomes and LE awareness.

This project aims to acquire a supercomputer cluster that will in turn enable data-intensive research in many diverse fields such as bioscience, geoscience, imaging, and vision. Specifically, the acquisition responds to the need to transform campuses' supercomputer resource provisioning practices with federated, hybrid cloud services that can seamlessly orchestrate the provisioning of local and remote resources (e.g., cyber-enabled scientific instruments, public clouds) to meet data-intensive research and education needs of users. The project focuses on application workflows considering connectivity and communications necessary for interdisciplinary research and education collaborations.

The supercomputer cluster augments existing facilities (e.g., Science DMZ 'network instrument' connected to the Internet2 Innovation platform, Transmission Electron Microscope, Federated-IAM 'entitlement service') and the core on-campus supercomputer resources. The project leverages advanced tools for central processing units and graphics processing units (CPU/GPU) as well as network visualization from existing funded projects for configuration of experiments with the instrument to develop 'custom templates' for diverse data intensive web-based applications. These custom templates will abstract the high-level policy and performance throughput requirements of data-intensive applications and 'personalize' them to lower-level control specifications implementable in an on-demand manner by virtualization technologies such as OpenStack and OpenFlow. Furthermore, the investigators will assess how the next-generation supercomputing user service models with custom templates can be composed to allow campus IT staff to sustainably and seamlessly support hybrid cloud use cases in research and education. Specifically, the instrument services
- Hybrid Cloud Computing,
- Bioinformatics and computational biology,
- Multi-modal data analytics, and
- Next generation HPC user services.

The instrumentation supports 16 researchers and their external collaborators in diverse data-intensive science fields such as bioscience, geoscience, imaging, and vision. It also supports the delivery of high-performance computing and Big Data analytics courses to more than 500 students at this institution and those around it. Participation of underrepresented and underserved groups will be accomplished utilizing the current NSF REU site as well as the institution's EPSCoR activities. Best practices to streamline the engineering/operations of hybrid clouds for data intensive applications, technologies/tools, policies and service models will all be disseminated. The instrument supports undergraduate and graduate courses and various other education and training activities, including REU programs. The proposed projects advances computational science research, research training, and curriculum development.

This REU Site award to the University of Missouri (MU), located in Columbia, MO, will support the training of 12 students for 9 weeks during the summers of 2014-2016. This award is supported by both the Directorates for Biological Sciences (Division of Biological Infrastructure) and Mathematical and Physical Sciences (Division of Mathematical Sciences). Pairs of student researchers, one from the computational sciences and the other from the life sciences, will work collaboratively at the interface of high performance computing and big data biology. Students will select from a diverse group of research projects, with a choice of mentors from five departments in the computational and life sciences. Example projects include developing parallel computing approaches to allow the identification of gene duplications that occurred with important evolutionary transitions such as the origins of mammalian placentation and using parallel computers and GPUs to infer the plant tree of life with data from high-throughput sequencing. In addition, participants will join other campus summer research programs -- attending professional development sessions, training in the responsible conduct of research, and participating in weekly sessions on parallel computing and biology. Participants will be recruited from a variety of institutions, with preference given to students from minority-serving institutions and undergraduate-only colleges. The PIs and research mentors will select the participants (half from the computational and half from the life sciences) and help them choose a computational or biological partner. Participants' progress will be assessed with pre- and post-tests covering the program's technical material. The program will be evaluated using NSF's common assessment tool.

Computation, and especially parallel computation, will be vital to science and engineering in the future, as increasing data volumes from instruments such as next-generation DNA sequencers generate enormous amount of data. The training and research experience offered at this REU site will give the participants unique opportunities to be trained in this important area.

Students are required to be tracked after the program and must respond to an automatic email sent via the NSF reporting system. More information is available by visiting http://muii.missouri.edu/reu_bigdata, or by contacting the PI (Dr. Gavin Conant; conantg@missouri.edu) or the co-PI (Dr. Chi-Ren Shyu; shyuc@missouri.edu).

? DESCRIPTION (provided by applicant): To train the next generation of scientists and informaticians, the University of Missouri Informatics Institute, teaming up with 23 faculty from Veterinary Medicine, Human Medicine, Engineering, Animal Sciences, Statistics, Nursing, and Journalism, has developed a collaborative pre-doctoral training program to provide elite incoming students with specialized training in Big Data under the One Health theme. This proposal requests support from the National Institutes of Health for a training grant on Massive and Complex Data Analytics, Pre-Doctoral Training in One Health. This training grant will provide funding to support six trainees per year. The faculty members participating in this training program comprise an exceptional group of outstanding scientists selected from MU's highly-interdisciplinary pool of researchers and educators. Students will participate in a training program that has both department-specific and training program-wide components. Our unique departmental components include: (1) Required Data Science and Analytic classes that are highly personalized to ensure core competencies for trainees from diverse technical backgrounds. (2) A specialized, tailored informatics curriculum appropriate for the unique research interests of each trainee; (3) A group of outstanding scientists to serve as academic mentors and research role-models for intra- and interdisciplinary research; Our interdisciplinary components include: (1) Tri-labs rotations, allowing students to gain hands-on wet-lab experience in human and animal health, as well as informatics; (2) Instruction in written and oral communication, so that our trainees will be able to efficiently disseminate their analytics outcomes for actionable plans; (3) A suite of creativity events and student-driven seminars, allowing students to work in teams with senior researcher to address and identify current and future research challenges; and (4) Professional networking that enables our trainees to present their work at national and international meetings. Together, these departmental and program-wide components provide our trainees with a depth of disciplinary expertise, and a breadth of exposure to other disciplines. Under the One Health theme, a new breed of data analyst will be trained; one who can quickly analyze animal and human data and infer discoveries for improved human health.

The rapid global spread of COVID-19 is changing the way we handle pandemics and has prompted academia, private industry, and government to leverage new technologies and approaches to minimize the impact on human lives. While much attention is on outbreaks in urban areas, the efforts to assist rural populations has lagged. The rural communities encompass roughly 19.3% of the US population and 95% of the US land area. They are especially vulnerable to disease outbreaks due to lower levels of necessary resources, such as access to hospitals, internet, 911, as well as overall lower socioeconomic status. The impacts of COVID-19 in rural areas are expected to be devastating. This project delivers research, scientific, and COVID-19 planning to three rural communities. The deliverable to the research community is access to hundreds of layers of integrated geospatial data that are available for advanced queries and visualization of results to support their own COVID-19 research. In addition, research results will enhance the understanding of disease transmission behavior and enable preparation for resilience in rural populations. The scientific community will receive new computational methods inspired by the rural disease analysis and associated resource management need assessment and tracking. The implementation of this mathematical and computational work will be made available to the COVID-19 planning community, including rural stakeholders, by creation of an interactive dashboard where maps and summaries will provide the frontline clinicians and/or public health responders up-to-date reports and context for specific rural areas. The project focuses on Missouri?s rural areas with a plan to extend the framework to the bordering states.

This project addresses whether the recent advancements in geospatial and network analyses can be leveraged to provide a scalable connected health ecosystem for rural America in response to the COVID-19 outbreak. It also address the new innovations necessary to bring explainable intelligence the future waves of COVID-19 outbreak. To answer these issues, the research team, consisting of experts in computing, geoinformatics, influenza, virology, pathology, acute care, and telemedicine, plans the following: (1) the team will first rapidly extend their previous work with the unique GeoSPatial Analytical Research Knowledgebase (GeoSPARK) big data framework with relevant data from the Census, healthcare systems, as well as the evolving information surrounding COVID-19 disease dynamics. GeoSPARK will provide real-time analysis using advanced complex queries across multi-resolution locational information to address the lack of an integrated data framework dedicated to COVID-19 risk assessment, capacity investigation, and geo-enabled decision support. (2) The team will develop and implement a suite of geospatial analytic methods which are inspired by the dynamics of disease outbreaks, such as network analysis (e.g., scenario analyses ? analyze the sensitivity and impact of disruptions in resource distribution, containment, etc.), hot spot analysis, contextual analysis, clustering analysis, etc., to quantitatively weigh risk and assess the multi-faceted problem of rural disparity. The analytical tools and dashboards inspired by the field?s needs and disease dynamics in rural areas are transformative and will enable better understanding of scenarios other than COVID-19, such as zoonotic disease outbreaks, flooding, and earthquakes.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.