Tim is a Senior majoring in Civil and Environmental Engineering and minoring in Computer Science. Though he started research his Freshman year, it was not until the winter quarter of his Junior year that he learned of Dr. Maya Cakmak’s human-centered robotics lab – the place he now calls home. Tim is excited about the field of human robot interaction. He programs robots to work with and serve society, allowing us to accomplish significantly more than we could ever do alone. His current project involves developing a web-based interface for robot teleoperation which allows for efficient and accessible control of a robot to perform mobile manipulation tasks. Tim loves the complexity of robotics, and the way it brings people together. He wants to contribute to the field in such a way that people of all backgrounds and experiences can benefit from the proliferation of robots. Whether this involves making teleoperation more accessible to people with severe physical disabilities, or developing applications that allow non-roboticists to program robots to complete meaningful tasks, Tim works diligently to make robots as helpful as possible. When not found in the lab writing code for the Fetch or PR2, you can likely find Tim spending time with friends from his church community, or studying for the next exam. After completing his last year of undergraduate studies, he plans to attend graduate school. Tim would like to thank the Washington Research Foundation for the honor of being selected for such a prestigious fellowship, and he would also like to thank Dr. Maya for her endless support and guidance in his research career.
Mentor: Maya Cakmak, Computer Science and Engineering
Project Title: Accessible Teleopertation of Robots for Mobile Manipulation Tasks
Abstract: Though robots have become common in many industrial facilities, they have not yet become widely adopted for domestic use. One of the challenges that has prevented the adoption of domestic robots is a lack of accessible, easy-to-use teleoperation interface. Many teleoperation interfaces do exist, and some are specialized for specific tasks such as robotic surgery, or robotic bomb deactivation. The current standard interface used for mobile manipulation is RVIZ Interactive Markers which often take an experienced user over a minute to complete a simple task such as picking up a box and setting it down a few feet away. Such a task would only take a non-disabled person a matter of seconds to accomplish without using a robot. This sluggishness of the interface poses a major challenge for all users. If a disabled user wants to use a robot to clean up the house, retrieve an item, or open a door, the current interfaces available would make such tasks extremely time consuming. Therefore, my research project aims to create an accessible and easy-to-use interface that would allow a user of any ability to perform manipulation tasks in a matter of seconds. To achieve this goal, we plan to develop an array of teleoperation interfaces considering the most time-consuming aspects of the current interface, which were discovered through previous research. Novel users will then test our interfaces as well as the current standard to determine how they compare to each other and to the status-quo. Our interface will allow users of any physical ability to easily control a robot to perform important manipulation tasks in seconds, therefore increasing robot’s accessibility and their widespread adoption.
Timothy Bi is a senior majoring in Bioengineering. He has been working with Professor Valerie Daggett, who is also in the bioengineering department, since January of 2016. The Daggett lab primarily focuses on studying amyloid proteins via computational modeling and protein inhibitor design, as well as experimentally testing those designed inhibitors against amyloidogenic proteins. Timothy’s current research project involves characterizing the behavior of a synthetic, amyloidogenic peptide designed by the lab by measuring its aggregation and cell toxicity, the hallmark properties of any amyloidogenic protein. Timothy has also observed that the aggregation of this synthetic peptide is potently inhibited in a variety of naturally-occurring amyloidogenic peptides. By computationally modeling these interactions and corroborating his findings with experimental data, Timothy hopes to discover novel insights on what “triggers” the protein misfolding events that result in amyloid diseases. These novel insights can then be used to design effective amyloid inhibitors, which may eventually see clinical use for the treatment of a variety of amyloid diseases (including Alzheimer’s). Upon completion of his project, Timothy plans on attending medical school to pursue his passion of improving the quality of life for those who are terminally ill.
Mentor: Valerie Daggett, Bioengineering
Project Title: Using Designed Peptides to Identify Mechanisms of Amyloid Aggregation in Alzheimer’s and Design of a De Novo Inhibitor
Abstract: Alzheimer’s is a disease that affects millions of individuals worldwide. Currently, there are not many effective treatments available on the market that can slow the neurodegeneration of the disease. Recent evidence has shown that oligomers of beta-amyloid (Aβ) protein are primarily responsible for toxicity in the brain in Alzheimer’s, which has sparked research to develop inhibitors for oligomer formation. However, a critical problem is that the actual structure of Aβ oligomers is unknown, in large part because traditional methods used to determine protein structure are ineffective due to the transiency of these oligomers. There is a need to identify the structure of these oligomers to understand the mechanism of oligomer formation, which will lead to effective inhibitor design. A designed α-sheet peptide known as AP3 displays striking behavioral similarities to Aβ under low pH. In addition, AP3 aggregation and toxicity may be potently inhibited by other amyloid species. This interaction will be closely explored both experimentally and via computer simulation, and the data will be used to design a de novo peptide inhibitor. This inhibitor will then be tested for its ability to inhibit amyloid aggregation and toxicity, as well as whether it can specifically bind to Aβ in an extremely heterogeneous solution and low concentrations. Successful completion of this project will result in significant progress towards understanding amyloid behavior and aggregation mechanism, which may lead to further novel applications of the α-sheet structure in treatments for Alzheimer’s, including but not limited to the inhibitor designed in this project.
Rian Chandra is a 5th year student in the Interdisciplinary Honors program, majoring in Physics, and Applied Mathematics and Computational Sciences. He has worked for the past four years at the HIT-SI3 plasma physics lab, under Professor Tom Jarboe. The lab studies a novel technique to confine a plasma in steady state, for fusion power applications. His research has focused on optical diagnostics, primarily the lab’s Ion Doppler Spectrometer, which measures ion temperature and velocity in the plasma, through the Doppler broadening and shift respectively of impurity spectral lines. He has also lead a project to determine if computationally modeled spectral profiles, computed using precise knowledge of orbital electron interactions, could be used as a diagnostic tool. His current project builds off both these experiences: he will be working with computational models to apply laser stimulated emission spectroscopy to various classes of plasma jets. This project is in collaboration with Justus-Liebig Universität, Gießen. Experimental plasma research, as a frontier engineering-physics discipline is intrinsically interdisciplinary, and Rian has found that the creativity that questions in this research demand make each step of the process novel and exciting. Next year, he will be continuing this journey as a graduate researcher assistant.
Mentor: Brian Nelson, Electrical Engineering
Project Title: Investigations of a Collisional Radiative Model as a Diagnostic Tool for Electron Dynamics Inside a Plasma Jet
Abstract: Plasma jets have many important modern applications, from satellite technology to medical research. It is known that the electron dynamics play a key role in loss mitigation and control, but their accurate measurement has so far proven difficult. This project will attempt to evaluate the limits of one possible diagnostic solution, the “Collisional Radiative Model” (CRM), by application to underexplored parameter regimes. Plasma parameters are inferred by comparing observed spectral intensities to a model, which in turn relies on accurate knowledge of orbital electron interaction cross sections across a wide range of input parameters (in our case, electron temperature and density). This project will work in parallel with a group at JLU-Giessen, who have developed a CRM based on the LXcat atomic database, and will be applying it to a low temperature Dielectric Breakdown Discharge ion thruster (r=1cm). A new model will be built using different atomic data (from the ADAS dataset) for verification, before being applied to two jets with atmospheric gas and pressure conditions (a Hall thruster (r=4cm) and a biomedical jet (r=1cm)). The model will include as many excited state population/depopulation mechanisms as possible, including electron and ion impact, recombination, and ionization. Spectral data will be collected with a Princeton Instruments Intensified-CCD fast camera optically coupled to a grating spectrometer. The goal is to probe the limits of application of the CRM by attempting to make high spatiotemporal resolution maps of the electron dynamics, in this new parameter regime. If time permits, the significance of nonequilibrium effects will be investigated.
Micaela Everitt is a senior at the University of Washington studying Bioengineering. During her freshman year, she joined the Musculoskeletal Systems Biology Lab in the Orthopaedics and Sports Medicine Department with Dr. Ronald Kwon. The lab studies the bone regeneration process in zebrafish in hopes to translate the research to a human model. She began to research a way to produce time-lapse images of the zebrafish regeneration process in order to understand the dynamic process. With the support of this fellowship, she has begun to design and construct an imaging chamber for this application. The bone regeneration process in zebrafish is similar to the bone development process in humans. Thus, a better understanding of the zebrafish regeneration process could aid in the development of digit or limb regeneration therapies in humans. In addition to her research, she is co-project manager for a Bioengineers Without Borders team. The team is designing a point-of- care hydration monitor for developing countries. Micaela also spends a lot of her time tutoring math at CLUE. After graduating in the spring, she hopes to continue on with research by pursuing a PhD in Bioengineering. She hopes to follow her passion for research and someday be the principal investigator of her own lab. Micaela would like to acknowledge all of the support from her mentor Dr. Ronald Kwon. She would also like to express her gratitude to the Washington Research Foundation for supporting her research endeavors.
Mentor: Ron Kwon, Orthopaedics and Sports Medicine
Project Title: Engineering a Long-Term, Anesthesia-Free System for Imaging Regenerating Zebrafish Tailfins
Abstract: Currently, in order to image the bone regeneration in zebrafish, the fish need to be immobilized with the use of anesthesia. However, the standard for anesthesia currently prevents regeneration when used over a long period of time. Due to this, most lab protocols say that a zebrafish can only be dosed and imaged once per day in order to prevent this regeneration. There is a need for a way to produce time-lapse images of the zebrafish regeneration process in order to understand the dynamic process that currently can only be seen in 24-hour snapshots. The Musculoskeletal Systems Biology Lab (MSBL) proposes to use a 3-D printed mechanical restraint instead of anesthesia in order to restrain the zebrafish. This should circumvent this lack of regeneration problem and allow us to image the zebrafish for a period of up to 24-hours at a time in order to see the dynamic bone regeneration process. We believe that the bone regeneration process in zebrafish in similar to the bone development process in humans, thus a better understanding of the zebrafish regeneration process could aid in the development of digit or limb regeneration therapies in humans.
Sedona is currently a senior majoring in Neurobiology and Biochemistry at the University of Washington. She has known since high school that pursuing scientific research is her passion, and she has carried out research in several different settings during her time as an undergraduate student. Sedona started doing research during her first year at the University of Washington by joining the Pallanck Laboratory in the Department of Genome Sciences, where her work focused on studying the mechanism whereby mutations in a particular lipid metabolic gene cause increased risk for Parkinson’s disease using Drosophila as a model organism. During the summer following her junior year, she carried out a research project at the École Polytechnique Fédérale de Lausanne in Switzerland which laid the groundwork for establishing a Drosophila model system to study the pathology of the neurodegenerative disease-associated protein Tau in neuralcircuits. Most recently, Sedona’s interest in the relationship between the central nervous system and the gut microbiota has led her to join the Palmiter Laboratory in the Department of Biochemistry. In the Palmiter Laboratory, she is investigating whether a group of neurons in the hypothalamus which are known to drive feeding behavior, AgRP neurons, may also have a role in shaping the composition of the gut microbiota to promote energy homeostasis. This research project, which has generously been awarded funding by the Washington Research Foundation, could provide insight into how humans regulate the microbiome in health and in disease. Following the completion of her bachelor’s degree, Sedona plans to attend graduate school in order to earn a Ph.D. and pursue a career in academic research. Outside of the lab, Sedona enjoys playing guitar, baking, traveling, and being engaged in service and community through her church.
Mentor: Richard Palmiter, Biochemistry
Project Title: Investigating whether AgRP Neuronal Maintenance of Energy Homeostasis in Mice Involves Regulation of the Gut Microbiota
Abstract: The gut microbiota plays a critical role in host energy homeostasis. Since energy homeostasis is essential for survival, it is probable that structures involved in maintaining host energy homeostasis would therefore evolve to have mechanisms for shaping the composition of the gut microbiota to modulate its energy harvest capacity in alignment with the energy needs of the host. In humans and mice, an important structure in maintenance of energy homeostasis is a subset of neurons in the arcuate nucleus of the hypothalamus which coexpresses agouti-related protein and neuropeptide Y and which balances energy deficiency by driving feeding behavior (AgRP neurons). I hypothesize that activation of AgRP neurons also balances energy deficiency by altering the composition of the gut microbiota to a state with a greater capacity for energy harvest. To test this hypothesis, I will express an excitatory designer receptor exclusively activated by designer drug (DREADD), the hM3Dq receptor, in the AgRP neurons of mice and analyze the composition of the gut microbiota using 16S rRNA amplification, sequencing, and characterization following pharmacological activation of AgRP neurons in these mice (hM3Dq mice) relative to control mice (mCherry mice). I will assess the energy homeostatic effects of the altered gut microbiota by transplanting the gut microbiota of hM3Dq or mCherry mice into mice lacking native microbiota (germ-free mice) and measuring their subsequent bodyweight and energy harvest efficiency. I will then identify other brain structures involved in the neural circuitry underlying AgRP-mediated alteration of gut microbiota composition by activating only the AgRP neurons which project to two target structures of interest, the paraventricular hypothalamic nucleus and the lateral hypothalamic area suprafornical region, and assessing the subsequent gut microbiota composition and energy harvest capacity in these mice relative to control mice and mice with all AgRP neurons activated.
Rohan Hassan is a very dedicated senior who has always had a passion for science. His dedication stems from his pursuit for medical school. Even with commuting 20 hours a week, working in a research lab, and being a full time student he has maintained a strong GPA over the years. He is majoring in Molecular, Cellular, and Developmental Biology after which he plans to take a year off to work in a lab while volunteering in a clinical setting to save up money and strengthen his resume for medical school. After obtaining his MD, he plans on pursuing an MPH in order to become a physician who does clinical research. Currently his project focuses on the interactions between Helicobacter pylori and its host immune response. He is very thankful to his amazing lab members who were able to support him through the year especially his PI, Nina Salama, and his mentor Tina Gall. Furthermore, he is honored and ecstatic to be one of the few that were accepted for the Washington Research Foundation Fellowship and would like to thank the WRF for their support.
Mentor: Nina Salama, Microbiology
Project Title: Characterizing the Role of LCN2 during the Gastric Epithelial Cell Immune Response Initiated by Helicobacter pylori
Abstract: Dr. Salama’s lab studies Helicobacter pylori, a gastric bacterial pathogen that infects about 50% of the world’s population. Chronic infection with H. pylori causes inflammation and increases the risk for developing gastric cancer. H. pylori colonizes the human stomach where it uses a type IV secretion system to deliver bacterial factors to the gastric epithelial cells. Once infected, gastric epithelial cells initiate different immune responses, one of which is the release of a protein; lipocalin 2 (LCN2). Like nearly all pathogenic bacteria, H. pylori must acquire iron, which is important for colonization, persistence, and virulence. LCN2 takes advantage of this fact and behaves like a competitor for iron essentially preventing bacterial growth by depleting intracellular iron stores. LCN2 is highly upregulated in gastric tissue during H. pylori infection, however it is still unclear if LCN2 plays a beneficial role for the host by controlling H. pylori growth. My project will test the hypothesis that if host LCN2 is sequestering iron then H. pylori survival rates will be decreased. The first aim will be to characterize the kinetics of LCN2 expression determining when the protein is most highly expressed during H. pylori infection. The second aim will use novel and innovative CRISPR/Cas9 genome editing techniques to engineer LCN2 knockout gastric cells. I can take these cells and co-culture them with H. pylori. If my hypothesis is correct, I expect that H. pylori survival rates will increase when co-cultured with LCN2 knockout cells since there will be no sequestering of iron by LCN2. Understanding how the host tries to defend itself through proteins such as LCN2 will provide better insight into the host immune response, promoting vaccine design and novel treatment strategies targeted to those populations with a higher predisposition to developing gastric cancer as a result of H. pylori infection.
Jessica Johnson is a senior majoring in Bioengineering. With an interest in neuroscience and translating her bioengineering curriculum to practical applications, she joined Dr. Rajiv Saigal’s lab in the fall of her junior year. As Dr. Saigal is a clinician, he has helped Jessica to understand the clinical need for a product and prioritize translational research. With the support of the Washington Research Foundation fellowship, she is planning to develop a hydrogel-based controlled drug release system for the treatment of spinal cord injuries. In addition to research, Jessica is the undergraduate student lead for BioEngage, an organization within the Bioengineering department that helps connect students to industry. After graduating, Jessica intends to pursue a Ph.D. in Bioengineering, with a focus on neuroscience applications, to enter the field of medical product design and development. She would like to thank the Washington Research Foundation for awarding her the fellowship and supporting her research.
Mentor: Rajiv Saigal, Neurological Surgery
Project Title: Controlled Drug Release for the Treatment of Spinal Cord Injuries
Abstract: Spinal cord injuries affect over 276,000 people in the U.S. alone. After the initial injury, a secondary injury occurs which results in further neuronal cell death as a result of the inflammatory response and microglial activation in the damaged tissue. Anti-inflammatory drugs have proven effective in reducing the secondary response, however side effects prevent their full-scale clinical use. A localized, controlled system is needed to target the secondary response and reduce further cell death. To solve this, we will design a controlled, localized drug release hydrogel-based system that can be placed on the spinal cord injury. The localized system will allow for a lower steroid dose, thus side effects should be reduced, and the controlled release should improve the efficacy of the drug. With the continued damage reduced, patients should see increased mobility.
Julia is a senior studying Biochemistry and Molecular, Cellular, and Developmental Biology. She is exploring the mutator phenotype of cancer, which theorizes that mutation rates are elevated in cancer cells and thus lead to the accumulation of mutations that accelerate tumorigenesis. She is interested in identifying mutations in genes that can modulate mutation rates, as understanding their functions may contribute to novel approaches for cancer treatments by targeting the mutator phenotype. She is specifically investigating the methods by which Chromosome Transmission Factor (Ctf18) suppresses mutation rates in yeast deficient in DNA polymerase epsilon proofreading, potentially via novel mechanisms involving direct interaction with the polymerase. Outside of the lab, Julia enjoys singing in UW Chorale and helping other students to realize their academic and professional goals as a CLUE Chemistry tutor and Undergraduate Research Leader. She is incredibly grateful to the Washington Research Foundation for their support of her growth as a scientist.
Mentor: Alan Herr, Pathology
Project Title: Assessing the Role of Ctf18 in Aiding and Abetting Mutator Polymerases
Abstract: Mutator phenotypes due to mutations in genes encoding DNA polymerases or mismatch repair proteins lead to increased error rates during DNA replication that accelerate the evolution of cancer cells and contribute to chemotherapy resistance. Work in the yeast Saccharomyces cerevisiae indicates that excessive DNA replication errors can lead to “error-induced extinction” (EEX), where every cell within the population dies due to a random lethal mutation. Thus, modulating mutation rates of mutator cells, to either exacerbate or suppress mutation rates, represents a possible direction for cancer therapy. In a recent screen isolating “antimutator” mutations that rescue cells from EEX by suppressing mutation rates, our lab identified a mutation in Chromosome Transmission Fidelity 18 (ctf18-K666fs). Ctf18 plays multiple roles in the cell that may influence mutation rate. My objective in this study is to identify which of its functions contribute to its antimutator effect. My first aim is to test whether the known role of Ctf18 in activating the S-phase checkpoint pathway is sufficient to explain the antimutator phenotype of ctf18-K666fs. To accomplish this, I will compare the mutation rates of dun1∆ ctf18-K666fs and dun1∆ ctf18∆ to their respective single mutants in cells expressing the pol2-L439V allele, which is defective in polymerase proofreading. Lower mutation rates in the dun1∆ ctf18∆ double mutants would indicate additive antimutator effects from independent pathways. My second aim is to determine whether the other functions of Ctf18 — PCNA loading and unloading and cohesin stabilization—contribute to the pol2-L439V mutator phenotype. I will investigate these potentially novel antimutator mechanisms using separation of function mutations and comparing mutation rates of mutant and wild type cells. Expanding our understanding of mechanisms by which antimutators can modulate mutation rates may contribute to novel approaches for cancer treatments by targeting the mutator phenotype.
Jacob Kowalsky is a senior studying microbiology with a focus on pathogens and pathogenesis. For the past two years, he has been pursuing this interest by studying influenza under the mentorship of Dr. Jesse Bloom and Dr. Alistair Russell in the Bloom lab at the Fred Hutch Cancer Research Center. With the support of the WRF Fellowship, Jacob is researching the interaction between influenza particles and host cells. Specifically, his project studies the mechanisms through which host cells recognize they are infected as they begin the process of fighting the virus with a host immune response. During the course of this project, custom flu particles containing mutated proteins and fluorescent protein markers will be used in combination with antibody staining of host cells in order to build a mechanism for how influenza uses viral components to suppress the immune response of the host. Understanding how certain flu particles trigger an immune response could also be utilized to improve influenza vaccines by triggering a stronger innate immune response within the subject. After graduation, Jacob plans to work as a laboratory technician in a virology lab before attending medical school. Jacob hopes to apply his medical degree to a field related to his research by studying how diseases spread in developing communities both within the United States and abroad.
Mentors: Jesse Bloom, Division of Basic Sciences; Alistair Russell, Division of Basic Sciences
Title: Internal Deletion Induced Interferon Response to Influenza A
Abstract: The innate immune system serves as a key first line of defense against infection by the influenza A virus, with the signaling components, called interferons, driving the production of a potent cellular antiviral response. Studies have indicated that influenza populations replete in defective virus particles, virions with a deletion in a portion of their genome, are less efficient at blocking the antiviral response, as shown by increased interferon in the host. Our project seeks to explore this phenomenon of RNA deletions leading to increased interferon expression in host cells by testing the hypothesis that deletions in the three polymerase genes of influenza alone are sufficient to cause an increase in the interferon response. To begin answering this question, an interferon reporter system was used to analyze the viral genome of interferon positive cells, and deletions within various lengths of the PB1, PB2, and PA polymerase genes were found to be enriched. Using these enriched sequences, my mentor, Dr. Alistair Russell, created pure populations of PB1 and PB2 defective influenza and found that they were capable of inducing interferon. I then supported this study by creating PA expressing cell lines and influenza with deletions in PA, with the objective of determining whether influenza particles with deletions within PA participate in the same phenomenon and are sufficient to induce interferon. My next endeavor is to create influenza with a fluorescent protein tag attributed to the NS viral protein. When mixed with polymerase defective particles to co-infect cells, this strain will allow observations to be made regarding the frequency of NS absence in the context of induction by defective particles. It is hoped that these results will support efforts to explain how influenza interacts with the immune system and is capable of subverting the host antiviral response.
Briana Lee is a senior undergraduate student majoring in both neurobiology and biochemistry. Briana is interested in how physics and chemistry can be applied to explain biological phenomenon. She has a passion for interdisciplinary studies and her current research involves processing neuroimaging data to analyze functional connectivity networks in Alzheimer’s Disease (AD). Functional networks are clusters of neurons within the brain that activate together when completing a certain task. These networks are not well characterized or understood and Briana is interested in quantifying how a specific functional network called the default mode network (DMN) changes during AD progression. She is also interested in how the DMN behaves differently across individuals, potentially providing insight to how environmental or demographic characteristics can affect an individual’s network capabilities. Through this work, with the support of the Washington Research Foundation Fellowship, Briana is excited to fully pursue her research project and is looking forward to learning as much as she can in the process. After graduating from UW, she hopes to attend medical school, while continuing research in a neurobiology-related field. Apart from research, Briana enjoys long distance running, reading, and cooking. Briana would like to thank Tara Madhyastha for her incredible support, she could not have asked for a better mentor. She would also like to thank the Washington Research Foundation for giving her the opportunity to pursue her passion for research and learning.
Mentor: Tara Madhyastha, Radiology
Project Title: Longitudinal Change in Default Mode Network Connectivity and Memory Scores in Alzheimer’s Disease
Abstract: During the long preclinical phase of Alzheimer’s Disease (AD), patients with similar levels of brain pathology may display differing levels of cognitive impairment. We propose to study potential physiological reasons for this variability. Alzheimer’s disease is a neurodegenerative disease characterized by destruction of temporal lobe structures, specifically the hippocampus. The hippocampus plays a crucial role in memory function, and implicated in a functional brain network called the default mode network (DMN). The DMN is a collection of cortical regions that work together, and are active at rest. Because this network is active at rest, functional magnetic resonance imaging (fMRI) scans taken while a subject is resting quietly may be used to analyze DMN activity. In our study, we will investigate how the DMN connectivity relates to cognition in subjects with various levels of impairment. We will obtain neuroimaging data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database from three subject pools: healthy controls, mild cognitive impairment, and Alzheimer’s disease. We will use neuroimaging analysis methods to calculate the connectivity within the DMN. We will use a parallel process model in R, using software developed in our lab to quantify the relationships between change in memory and DMN connectivity. We intend to present our results at the Alzheimer’s Association International Conference (AAIC) in summer of 2018.
Jason is a fifth year senior at the University of Washington in the department of Bioengineering. His interests in brain injury led him to join Dr. Pun’s lab to study methods of drug delivery to the spinal cord. His current research focuses on peptides that can self-assemble into hydrogel structures. These hydrogels can be functionalized to deliver drugs as well as to facilitate stem cell growth. The ultimate goal of his research is to create a hydrogel formulation that can be injected into the spinal cord to combat the effects of traumatic spinal cord injury. After graduation, Jason plans to pursue medical school in order to conduct research in academic medicine. He would like to thank his mentors Dr. Suzie Pun and Dr. Tianyu Zhao for their guidance and encouragement throughout his research endeavors. He would also like to express his gratitude to the Washington Research Foundation for their generous support of his project.
Mentor: Suzie Pun, Bioengineering
Project Title: A Self-Assembling Injectable Hydrogel Formulation for Spinal Cord Rehabilitation
Abstract: Spinal cord injuries are particularly devastating because there are no clinically effective methods to regenerate damaged neural cells, and the inflammatory response can cause secondary injuries and inhibit regeneration. Novel treatments like cell therapy and drug delivery have been found to have greater efficacy when delivered by hydrogel scaffolds. Currently, there is no clinically effective way to therapeutically scaffold cells and deliver thrombin inhibiting drugs into the spinal cord. Therefore, there is a need for an effective treatment of spinal cord injury that can locally deliver therapies with anti-inflammatory properties to the spinal cord, as well as provide a support material for neural stem cell therapy. We propose a hydrogel composed of self-assembling peptides appended with peptide sequences to have bioactive properties. These peptide sequences will include motifs associated with the extracellular matrix to promote cell growth, and anti-inflammatory drugs to combat thrombin upregulation during inflammation. The peptides will be kept in solution, mixed with neural stem cells, and injected into the spinal cord where the physiological pH and ionic salt conditions will catalyze formulation of the gel. This will ultimately work to provide both effective cell therapy and anti-inflammatory drug delivery to the spinal cord, with the potential for broader applications in the future.
Nelson Liu is a junior majoring in computer science, linguistics, and statistics. His research interests lie in the fields of natural language processing and machine learning. He started working with Professor Noah Smith from the onset of his first quarter at UW. Through his work with Professor Smith and research internships, Nelson has been fortunate to explore subfields such as computational social science, question answering, and machine translation. With the support of a Washington Research Foundation Fellowship, Nelson seeks to characterize the limitations of recurrent neural networks, a popular model for natural language and sequential data in general. After completing his undergraduate degree, Nelson plans to pursue a Ph.D. in natural language processing and finally a career in research.
Mentor: Noah Smith, Computer Science & Engineering
Project Title: Examining the Ability of Recurrent Neural Networks to Handle Long Range Dependencies
Abstract: Deep learning has pushed the boundaries of natural language processing, and neural network architectures have achieved state of the art results in tasks such as automatic machine translation, question answering, and language modeling, among others. The backbone of many of these models is the recurrent neural network (RNN), which serves as a general-purpose module for encoding sequences of text as a vector of real numbers. Despite their ubiquity, RNNs are often treated as black box text encoders, and little is understood about the information stored in their internal learned representations. RNNs have become popular for natural language processing because in theory, they are capable of considering all previously seen information when solving a task at hand. For example, this ability is critical in language modeling (the task of predicting the next word in a sequence) because the next word is often heavily dependent on prior context. However, in practice, RNNs don’t seem to connect these long-term dependencies inherent in text. In this work, our goal is to determine the extent that RNNs remember earlier parts of a text sequence. We propose to open the black box through a principled set of experiments for empirically evaluating the memory capacity of RNNs. We will examine whether RNNs can hold and use information indefinitely, or whether they have a fixed memory length. Furthermore, we will investigate whether potential memory constraints arise from inherent limitations of the model, or the way in which it is trained. If the latter, we will explore new methods for more effective training of RNNs. By gaining a deeper understanding of the internal behavior and shortcomings of RNNs, we will enable the community to work toward overcoming these limitations and facilitate the design of faster, parameter-efficient RNN variants.
Caleb Perez is a fourth-year undergraduate student in the Department of Bioengineering. Since his junior year of high school, he has been deeply involved in biomedical research, driven by a commitment to advancing the field of medicine. His current work, in Dr. Suzie Pun’s lab, aims to improve one of the most promising recent innovations in cancer therapy: chimeric antigen receptor (CAR) T cells. Outside of the lab, Caleb participates in Bioengineers Without Borders, working on the design of a low-cost hydration monitor for application in the developing world. He also helps prospective bioengineers through their coursework as an undergraduate teaching assistant for three departmental classes. Following his graduation in the spring, he plans to continue biomedical research through graduate school and an eventual career in translational cancer research. Caleb would like to express his utmost gratitude to his mentors, Dr. Suzie Pun and Brynn Olden, without which his research would not be possible. He would also like to thank the Washington Research Foundation for its generous support.
Mentor: Suzie Pun, Bioengineering
Project Title: Cell-Molded Silica Microparticles as Artificial T Cell Activation Platform
Abstract: T cell-based immunotherapy has shown immense therapeutic potential for the late-stage treatment of cancer. An important step in the generation of these therapies is the ex vivo activation of T cells, which is generally done with the use of artificial antigen presenting cells (aAPCs). Although intended to mimic the function of immune cells in the body, currently used aAPCs do not fully recapitulate natural size, morphology, or membrane fluidity, all of which have been demonstrated to be important determinants of activation properties. Therefore, there is a need for an activation platform that more closely mimics antigen presenting cells (APCs) in vivo, which would likely significantly improve the efficiency of T cell activation for use in cancer immunotherapies. To achieve this, we propose using cell/silica composites that retain the size and morphology of their cellular templates. The fusion of an antibody-loaded lipid bilayer to these silica microparticles will then result in a closer recreation of natural APCs. We will optimize the fabrication of these novel aAPCs and fully characterize their activation properties. If these results indicate that this platform successfully improves the efficiency of T cell activation, this has the potential to significantly improve the cost and scalability of these promising cancer treatments.
Kimberly Ruth is a junior double majoring in Computer Engineering and Mathematics; she is also in the Interdisciplinary Honors Program and is pursuing Departmental Honors. Her research interests lie within the broad area of computer security, aiming to make computer systems stronger by understanding their weaknesses. Since winter quarter of her freshman year, she has been an undergraduate researcher in the CSE Security and Privacy Lab, co-advised by Professors Franziska Roesner and Tadayoshi Kohno. Her current research focus is on the security and privacy implications of emerging augmented reality (AR) technologies. AR systems present novel challenges for security due to their tight integration with the physical world, and Kimberly enjoys developing system design principles by analyzing these new risks; her current work aims to explore and re-define access control for multi-user AR systems, considering the implications of one AR user’s virtual content affecting another. To supplement her research work and gain a broad perspective on security and privacy, she participates in graduate-level security seminars, and last spring informally audited a course in cryptography. Kimberly maintains a parallel interest in mathematics and has participated in the Putnam competition. After graduation, Kimberly plans to pursue a PhD in computer science and subsequently a research-based career in computer security, leveraging mindsets of both theory and practice to inform the design of future secure systems. Kimberly is grateful to her fantastic advisors for their guidance and encouragement, and to the Washington Research Foundation for their support of her academic endeavors.
Mentor: Franziska Roesner, Computer Science and Engineering
Mentor: Yoshi Kohno, Computer Science and Engineering
Project Title: Secure Content Sharing for Multi-User Augmented Reality Applications
Abstract: Augmented reality (AR) technologies are rapidly emerging, from mobile applications to head-mounted displays to car windshields. Though there are many positive use cases projected for these technologies, the great power they have to affect a user’s perception of the world also comes with great risks. Prior work in security and privacy for AR has focused only on a single user interacting with a device. In my work, I examine emerging challenges with multi-user AR applications. These challenges are fundamentally different from those of existing technologies due to AR’s deep physical-world integration: content may be private or semi-public, for instance, in an inherently shared physical space, and virtual content access control might be determined in part by physical proximity. Drawing lessons from a preliminary user study that explored users’ expectations and worries about multi-user AR, I aim to characterize AR security and privacy challenges and, by means of a prototype, explore and evaluate solutions. I plan to describe a foundation for addressing these challenges in an academic paper, as well as to release a developer toolkit to enable real-world use of that foundation.
Riley Stockard is a junior majoring in bioengineering with a research interest in synthetic biology. She is working in the Klavins lab on a drug toxicity screen using genetically engineered yeast. The purpose of the screen is to vet drug candidates for off-target binding effects before they reach clinical trials, with the goal of creating safer drugs that reach the market faster. By engineering yeast to express proteins of interest on the surface of the cell and measuring the binding of cognate proteins, it is possible to represent an entire protein family in a single tube. Consequently, this screen can characterize the effects of a drug on hundreds of protein-protein interactions in a comprehensive and high-throughput way. Riley became involved in research her freshman year, which sparked her interest in synthetic biology and inspired her to apply to the bioengineering department. Post-graduation, she hopes to work in industry for some time before returning for a graduate degree. Riley is honored to be Washington Research Foundation fellow and would like to express her gratitude for the support she has received to continue her research endeavors.
Mentor: Eric Klavins, Electrical Engineering
Project Title: Reprogramming S. Cerevisiae for preclinical Alzheimer’s Drug Screening
Abstract: Alzheimer’s disease is a neurodegenerative disorder that is characterized by memory problems and eventual loss of cognitive ability. Although the cause of Alzheimer’s is unknown, the leading hypothesis since 1991 blames dysfunction in a network of protein-protein interactions centering around amyloid precursor protein (APP). No drugs targeting this network have been successfully developed, although they represent the majority of phase 2 and 3 clinical trials. Progress is obstructed by the laborious nature of drug development: 90% of drugs fail in clinical trials. Moreover, it takes over a decade for a new FDA-approved drug to reach the market and costs an average of $2.6 billion. Pharmaceutical companies are unable to comprehensively evaluate drug toxicity before clinical trials because there is currently no method to screen the thousands of protein-protein interactions (PPI) that a drug could unintentionally disrupt in the human body. The ability to disqualify drug candidates with off-target effects before clinical trials would greatly increase the speed at which useful drugs are introduced to the market and reduce the expense of developing new therapeutics for those who need them. Here, we aim to develop a library-on-library protein screen for Alzheimer’s drug candidates by utilizing synthetic agglutination of S. Cerevisiae that conserves the accuracy of pairwise toxicity screening in a high-throughput, one-pot format. Specifically, as an initial proof-of-principal, a ten-by-ten protein interaction network that includes proteins from the APP family will be constructed to probe the binding interactions within a complex and interconnected family of proteins. This interaction network will then be used as a tool to investigate drug candidates designed to act on the Alzheimer’s network.