Peng Wei, Ph.D. - US grants

DESCRIPTION (provided by applicant): An emerging research area in genetics is to detect associations between complex traits and rare variants (RVs) with next-generation sequencing data. Due to extremely low minor allele frequencies (MAFs) of RVs, many existing tests for common variants (CVs), such as the univariate test on each individual variant, most popular in genome-wide association studies (GWAS), may no longer be suitable. To boost statistical power, a common theme of existing association tests is to aggregate information across multiple RVs in a gene. With sequencing data, since the majority of RVs may not be causal, in which case most, if not all, existing association tests have severely deteriorating performance. We propose developing and evaluating an adaptive test that can maintain high power across various situations, including in the presence of opposite association directions and of many non-associated RVs. We will extend the proposed adaptive test to pathway analysis and multi-trait analysis. The developed methods will be applied to detect associations of RV-cardiovascular traits with the sequencing data from the CHARGE-S and ESP cohorts. We will develop and distribute software implementing the proposed methods.

DESCRIPTION (provided by applicant): Pancreatic cancer (PanC) is the fourth leading cause of cancer-related death for both men and women in the U.S. Better understanding of the etiology and developing risk prediction models for early detection and prevention are urgently needed for this rapidly fatal disease. The majority of PanC are caused by the interplay of both genetic and environmental factors. Known risk factors for PanC include cigarette smoking, obesity, long-term type II diabetes, and family history. In addition, our previous case-control study has shown that excess body mass index (BMI) in young adulthood confers a higher risk of PanC than weight gain at later age. Recent genome-wide association studies (GWAS) have identified several chromosomal regions and genes in association with risk of PanC (PanScan). Our pathway analyses of the PanScan GWAS data have uncovered several novel biological pathways associated with the risk for PanC. However, it remains unknown how environmental or host risk factors modify the association between genetic factors and the PanC risk, which knowledge is critical to better understanding of the etiology and developing a risk prediction model and early intervention strategies for PanC. The goal of this project is to identify gene-environment interactions and develop and validate a risk prediction model including both common and rare genetic variants using the PanScan GWAS data and the exposure information of over 2,200 case-control pairs and an ongoing ExomeChip-based study of PanC genotyping both common SNPs and >240,000 rare functional exonic variants in over 4,100 cases and 4,700 controls from six case-control studies in the Pancreatic Cancer Case Control Consortium (PanC4) and a nested case-control study from Europe (EPIC). We will validate the absolute risk prediction model in two large prospective cohorts: the Atherosclerosis Risk in Communities (ARIC) cohort of 15,000 individuals and the Kaiser Permanente cohort of 100,000 individuals. We will also develop novel statistical methods to identify genes modifying the association between changing BMI at different age periods and PanC risk using the unique dataset from a case-control study of PanC conducted at MD Anderson Cancer Center. Our proposed project hinges on novel integration of GWAS, ExomeChip, exposure data of a large number of PanC cases and controls, recently developed powerful statistical methods and analysis strategies for detecting genome-wide gene/pathway-environment interactions and polygenic approaches to genetic risk prediction. The work proposed here is expected not only to advance our understanding of the etiology of PanC and delineate how genes and lifestyle or host factors modify the risk of PanC, but also to greatly facilitate identification of high-risk individuals, and thus, contribute to early detection, improved survival and prevention of PanC. The novel statistical methods developed here are also applicable to other cancers and complex disease, and we will develop user-friendly software packages for public use.

DESCRIPTION (provided by applicant): This proposal is in response to PAR-13-382, supporting secondary data analyses of existing large genomic datasets for the purpose of identifying gene-by-environment (GxE) interactions. Lung function and its decline in older adulthood is likely the result of genetic and environmental influences. Cigarette smoking is a key environmental context for loss of lung function over time. Genome-wide association studies (GWAS) identified 26 genetic loci associated with cross-sectional spirometric measures of lung function. Recent GWAS of the longitudinal change in lung function have identified additional novel loci. To date, there is only one published genome-wide study of GxE interaction on lung function that considers smoking as the environment of interest. This genome-wide GxE study used common variation and cross-sectional information on lung function and smoking to identify three novel loci not previously associated with lung function. In aggregate, these published studies made important contributions to understanding the etiology of lung function, and were facilitated by the organizational structure and support of the Cohorts for Heart and Aging in Genomic Epidemiology (CHARGE) consortium and the CHARGE Pulmonary Working Group. Additional investigation is warranted to further understand how smoking interacts with genetic factors to influence lung function. The objective of this proposal is to elucidate the complex interplay of genes and environment underlying lung function using state-of-the-art statistical methods and analysis strategies that leverage available data resources. Ongoing work within the CHARGE Pulmonary Working Group includes analysis of data from the Illumina HumanExome BeadChip (the exome chip) for ~33,800 individuals of European ancestry with spirometric measures of lung function, all of whom also have longitudinal measures of smoking history and lung function. An additional ~6,000 individuals of African ancestry have measures of lung function, smoking history, and exome chip data, and ~3,800 also have longitudinal measures. Spirometric measures include forced expiratory volume in one second (FEV1), forced vital capacity (FVC), and their ratio (FEV1/FVC). These measures of lung function are important clinical tools for diagnosing pulmonary disease, classifying its severity, and evaluating its progression over time. The large volume of phenotype and exome chip data available within the CHARGE consortium provides a unique, cost-effective opportunity to apply new analytical approaches and methods. This application has two novel aspects: 1) investigation of rare variation and environmental interactions, and 2) investigation of longitudinal measures of environmental factors. The proposed research represents the next step in the efforts to investigate the interplay of genetic variation and environmental factors influencing lung function. Results from this study may disclose novel genetic susceptibilities to smoking exposure or a greater understanding of the role of smoking in the development, progression, and severity of declining lung function.

The objective of this project is to provide theoretical foundations for cyber-physical systems to support the increasing autonomy in the presence of other manned/unmanned air traffic. Currently the United States air transportation is facing significant challenges due to the rapid evolution of increasingly autonomous systems such as unmanned aircraft systems (UAS) and their expanding presence. However, there has been little scientific investigation on the cyber-physical systems that supports unmanned aircraft operations operating in the presence of other air vehicles. This research project explores novel strategies of coordinating and managing the UAS traffic to ensure low-altitude airspace safety and efficiency in near future. This will also help to catalyze additional cyber-physical systems research in related areas including aviation infrastructures, navigation and surveillance devices, and unmanned aerial vehicle technologies. The investigator will work with Iowa State University Extension and Outreach Office to promote UAS applications in precision agriculture, aviation safety and responsible uses of UAS.

The theoretical aspect of the research will leverage interdisciplinary methodologies in the fields of optimization, artificial intelligence, and control. The project designs algorithms, implements software and demonstrates proof-of-concept using low-altitude UAS traffic management as a challenge application area. This project establishes an unmanned air system traffic simulation platform, which can be used to validate new methods and tools, and enable collaborations with government, industrial and academic partners. The research outcomes are expected to provide a framework to investigate the feasibility, safety, efficiency and potential benefits of the increasing autonomy in civil aviation.

With the development of numerous civilian Unmanned Aerial System (UAS) applications, a large number of unmanned aircraft of various types need to be safely operated in low-altitude airspace. These UAS also need to safely share this space with manned aviation traffic, such as general aviation and helicopters. The FAA forecasts 7 million UAS sales (commercial and hobbyist combined) by 2020. This project advances and integrates the investigators' novel concepts of operations and core algorithms into an intelligent UAS Traffic Management system (UTM). This UTM system would also break market feasibility barriers for new UAS applications such as urban on-demand air transportation and UAS cargo delivery. Finally, the insights gained during the development of the proposed UTM could have profound impact on design and implementation of other human-centered smart service systems and cyber-physical systems that support civil aviation, e.g., air traffic infrastructures, operator ground support systems, communication, navigation and surveillance devices, and vehicle technologies.
This research proposes an intelligent UTM system integrating big data architecture and computation power that will coordinate pre-departure UAS flight plans, detect potential collisions in real time, generate recommendations to resolve potential conflictions, proactively control any risk to people and objects on the ground during an emergency landing, and identify the cause of collisions. The aim of these capabilities is to minimize the number of collisions and mitigate the impact of each accident. This will be achieved using large-scale optimization, aircraft guidance and control, predictive modeling, system verification and validation, and advanced visualization techniques for information presentation and decision support. The proposed system has a pre-departure flight plan coordination module that queries the approved flight plan database and performs conformance checking for every newly requested flight plan to achieve conflict-free pre-departure traffic coordination. An en route traffic monitoring and alerting module receives real-time aircraft position data and active flight plans, performs automated prediction for potential collision, and generate recommendations to resolve collisions. An emergency landing and contingency management module queries multiple databases such as terrain maps, obstacle data, airspace data, public safety data and real-time aircraft position data to suggest emergency landing site and calculate the corresponding landing path to minimize the impact risk to people and objects on the ground. Finally, the advanced human machine interface will provide information visualization and decision support in an intuitive way to reduce cognitive inefficiencies and maximize human-in-the-loop performance to augment UAS traffic controller capabilities. The proposed system will serve as a complementary component of an ongoing NASA UTM. The research plan has three phases: (Phase 1) Identification and synthesis of intelligent UTM user requirements, (Phase 2) Development of the intelligent UTM core algorithms and system prototype, and (Phase 3) intelligent UTM testing, evaluation, and integration.
This academe-industry partnership is lead by a multidisciplinary academic research team: Iowa State University (lead institution), University of Iowa (Iowa City, IA),and University of Michigan (Ann Arbor, MI),) with primary industrial partners Rockwell Collins (Cedar Rapids, IA) and Mosaic ATM (small business, Leesburg, VA) together with broader context partners the Federal Aviation Administration William J. Hughes Technical Center (FAA Tech Center) (government agency, Egg Harbor Township, NJ). The partners will also receive feedback from the FAA Iowa office and Uber Elevate. This partnership will ensure that the proposed UTM system meets FAA regulations, user requirements, and market needs.

The Quantum Leap Challenge Institutes (QLCI) program is part of the Quantum Leap, one of the research Big Ideas promoted by the National Science Foundation. This award is a QLCI Conceptualization Grant, which supports activities to build capacity among teams planning for the large-scale, interdisciplinary Challenge Institute projects that aim to advance the frontiers of quantum information science and engineering. Research at these Institutes will span the focus areas of quantum computation, quantum communication, quantum simulation, and/or quantum sensing. The Institutes are expected to foster multidisciplinary approaches to specific scientific, technological, and workforce development goals in these fields.

This Conceptualization Grant will develop well-formulated plans for a future Challenge Institute proposal along the theme of quantum photonics, topological computations, and molecular spintronics.

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.

The NSF Convergence Accelerator supports use-inspired, team-based, multidisciplinary efforts that address challenges of national importance and will produce deliverables of value to society in the near future. This project seeks to develop approaches that address issues of decoherence and crosstalk by scalable topological superconductors (TSC). Investigation for achieving realistic large-scale quantum computers is required to advance the field of quantum information science. This project will integrate Majorana zero modes (MZMs) into conventional superconducting qubit architectures to advance their application to quantum computing. By having outreach programs towards K-12 students and research experiences for undergraduates, this project will broaden participation in quantum with a focus on underrepresented minorities.

The project will build and establish a cross-sector team that will develop advances in controlling the topological nature of materials to advance quantum computing to deliver fault-tolerant qubits and their quantum interconnects. This project seeks (1) to understand and demonstrate the non-local topological nature of the MZMs by detecting the electron teleportation through a pair of MZMs; (2) to establish the basic elements for measuring the parity state in a teleportation-based T-qubit; (3) to explore flux quantization caused by a supercurrent loop that is mediated by the MZMs and set up the basic flux (or pseudo-spin) measurements of a T-qubit; (4) to identify and plan the Phase II research program, and (5) to build a strong team of academic, governmental lab, and industrial partners. Building on recent developments of a new TSC material platform, this project aims to demonstrate the quantum nature and the non-local topological protection of MZMs in the platform as well as build topological qubits that can be integrated into existing quantum computing circuitry. This may lead to greater functionality in superconducting circuits which can significantly advance topological quantum computing. The project deliverable includes a platform supporting topological qubit that is more robust and more scalable. By establishing a nationwide student exchange program and outreach activities to K-12 students, this project seeks to engage students in quantum research and training to broaden participation.

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.

Artificial Intelligence (AI) has rapidly become a critical domain with applications in autonomous driving, robotics, aerospace, healthcare, and others. For a machine (AI agent) to closely mimic human behavior and operate effectively, it should possess the capabilities of robust decision making and learning simultaneously as it operates in the environment. Multi-Agent Reinforcement Learning (MARL) is a promising research area that can model and control multiple distributed decision-making AI agents. However, recent studies have shown that the MARL algorithms suffer from inefficiencies that can severely limit their adoption in real-world systems. These problems occur due to complexities in decision-making processes arising from having to observe and act upon a large number of events present in the environment, along with the growth in the number of AI agents needed to interact with each other.

To ameliorate the learning efficiency and scalability issues of MARL algorithms, the project investigators adopt a novel interdisciplinary solution approach, harnessing computer architecture, machine-learning theory and optimization. Specifically, the project will seek techniques to improve neural-network throughput, to efficiently manage the state-action space in a dynamic fashion and to scalably encode states and observations of a large and varying number of agents. A hardware-software co-design approach is adopted to accelerate the concurrent optimization of software and hardware layers. The research outcomes of this project will significantly enhance the adoption of MARL frameworks in real-world applications and positively impact university curricular development and the computing industry.

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.