36214 research outputs found
Sort by
Improving Detection of Tuberculosis Disease Among Children
No full textThesis (Ph.D.)--University of Washington, 2025Tuberculosis (TB) remains a major public health concern globally and is a leading infectious cause of morbidity and mortality in both adults and children. Approximately 1.3 million children and young adolescents who developed TB disease in 2023, and among those 191,000 died from TB. The risk of TB infection progressing to disease, and from disease to death, is especially high in young children. Diagnosing TB in children is difficult, with an estimated 51% children globally with TB disease not diagnosed due to the lack of a quality specimen for testing, difficulties with specimen collection and the lack of highly sensitive tests for use in children. As a result, the majority of children with TB are empirically initiated on treatment based on their symptoms alone. This dissertation aims to address specific knowledge gaps and potential areas in the patient care cascade where gaps may occur to improve detection of TB in children. In my first aim, our objective was to evaluate the clinical diagnostic accuracy of lipoarabinomannan (LAM) in urine and whole blood samples in children with clinical TB symptoms using highly sensitive electrochemiluminescent immunoassays and the most accurate known LAM antibodies (FIND28:A194-01 & S4-20:A194-01). We used data from a prospective longitudinal cohort study of predominantly HIV-negative children with clinical TB symptoms recruited from inpatient and outpatient wards at Harry Gwala Regional Hospital in KwaZulu-Natal, South Africa. Using a composite reference standard for pediatric clinical TB diagnosis, we found that neither of the two current anti-LAM antibodies did not meet the WHO target product profile criteria for a non-sputum POC test (65% sensitivity, 98% specificity) in clinically diagnosed children. More sensitive and improved LAM detection antibodies are needed if LAM is to be a viable biomarker for diagnosing TB in children without advanced HIV disease.
In my second aim, we applied machine learning methods to identify clinical, biomarker and rapid diagnostic features predictive of unconfirmed TB in children in KwaZulu-Natal, South Africa. Using data from the previous study, we trained random forest models using kNN imputation and minority class oversampling strategies to predict TB status from a range of available features. Using mean decrease in accuracy and SHAP values to highlight model variable importance, we found that respiratory rate, serum C-reactive protein levels, asthmatic episodes, LAM positivity, admission for a lower respiratory tract infection and having a sick household member were offered predictive value in differentiating between unconfirmed and unlikely TB. Models incorporating inverse probability weighting to account for class imbalance prioritized medical history and exposure related predictors, while models using synthetic oversampling prioritized clinical and laboratory testing results for prediction. Including these accessible, non-sputum-based measures in pediatric TB treatment decision algorithms may improve diagnostic accuracy.
In my final aim, we used an agent-based model (TBSim) to estimate the effect of introducing novel diagnostics for both adults and children on pediatric TB outcomes in uMgungundlovu district, KwaZulu-Natal, South Africa. We evaluated four separate diagnostic strategies: (1) Xpert MTB/RIF Ultra alone; (2) Xpert plus oral swabs; (3) Xpert plus FujiLAM; and (4) Xpert plus a computer assisted detection for chest X-ray(CAD-CXR) for children. We found that when used as an add-on test to Xpert Ultra, oral swabs had the greatest impact on reducing pediatric TB incidence over the simulation period, especially in children 0-4 years of age. Surprisingly, none of the interventions substantially reduced overall TB mortality compared to Xpert Ultra alone. The inclusion of TPT prevention and social household contact network in future simulations are needed to verify findings.
Our findings illustrate that there are several opportunities to expand upon existing work in order to improve detection of TB disease in children. Improving detection, diagnosis and reporting of pediatric TB remains a complex challenge to address globally. Our findings highlight the need for child-centered approaches to developing novel TB diagnostics
Developing Feedback Systems For Live Stream Communities Using Large Language Models
Get PDFThesis (Ph.D.)--University of Washington, 2025Live streaming has become a key mode of community engagement, enabling real-time interaction between creators and audience. Platforms like Twitch have transformed content creation into a participatory, performative practice that blends entertainment, emotional labor, and community building. Despite its cultural and economic significance, little research has examined how streamers interpret and act on audience feedback in ways that support sustained growth, well-being, and meaningful interaction. This dissertation explores how large language models (LLMs) can power intelligent feedback systems that help streamers manage content production and community engagement. Through three studies, it examines feedback from multiple angles: streamers' existing information practices, proactive strategies for soliciting input, and reactive techniques for interpreting audience response post-stream. This research demonstrates how LLMs can synthesize audience input into insights that are actionable, emotionally supportive, and strategically meaningful. The work contributes both deployable systems and design principles for building feedback tools attuned to the demands of live stream communities
High-Performance Transaction Processing in Disk-based Databases
Get PDFThesis (Ph.D.)--University of Washington, 2025Achieving high-performance transaction processing in disk-based databases has long required system designers to choose between lock-based concurrency control methods, which suffer from CPU overhead and reduced parallelism, and timestamp-based methods, which provide superior concurrency but incur prohibitive I/O overhead when timestamp metadata is stored on disk. Modern high-speed storage devices like NVMe SSDs exacerbate this trade-off, as the CPU becomes the bottleneck for lock-based methods while disk-based timestamp storage wastes the storage device’s speed on frequent small metadata operations. This dissertation introduces a novel approach that eliminates this fundamental trade-off through approximate timestamp storage and demonstrates that timestamp-based concurrency control protocols—specifically Strict Timestamp Ordering (STO), Multi-Version Timestamp Ordering (MVTO), and TicToc—can maintain correctness (serializability) even when timestamps are overapproximated for inactive keys, as long as active keys maintain exact timestamps throughout their transaction lifetime. This key insight enables designing FPSketch, a hybrid data structure combining a hash table for exact timestamps of active keys with a probabilistic sketch for approximate upper bounds of inactive keys. The first contribution is the design, implementation, and evaluation of FPSketch integrated with STO, MVTO, and TicToc in the SplinterDB key-value store. FPSketch achieves nearly the idealized performance while requiring only minimal memory—as little as 32KiB for an 80GB database—by eliminating the need to access timestamp metadata from disk during normal operation. Experimental evaluation on modern NVMe SSDs demonstrates that TicToc with FPSketch achieves up to 14x higher goodput than traditional two-phase locking, up to 5.9x higher goodput than disk-based timestamp storage. The second contribution is a comprehensive analytical and experimental study evaluating FPSketch across the entire storage performance spectrum, from traditional hard disk drives with millisecond latencies to emerging CXL-based storage approaching DRAM-like speeds. The evaluation reveals that FPSketch’s benefits scale with the fundamental gap between local memory and remote storage access, ensuring its continued relevance as storage technology evolves. On slow storage (HDDs and SATA SSDs), FPSketch enables timestamp-based protocols to outperform traditional concurrency control methods: on SATA SSD, TicTocFocus-Sketch achieves up to 6.89× and 2.52× higher goodput than two-phase locking (2PL) and KR-OCC, respectively, while on HDD it reaches up to 1.8× the goodput of KR-OCC. FPSketch also eliminates the prohibitive overhead of timestamp disk accesses, achieving improvements of up to 569% over disk-based timestamp storage. On fast storage (simulated CXL-based SSDs), where systems transition from I/O-bound to CPU-bound, FPSketch continues to provide substantial benefits by keeping timestamp metadata in fast local memory, enabling timestamp-based protocols to significantly outperform traditional approaches. Together, these contributions establish that approximate, in-memory metadata management enables high-performance transaction processing for disk-based databases. FPSketch demonstrates that approximate metadata management can unlock advanced concurrency control designs that would otherwise be impractical, providing a practical solution that enables efficient timestamp-based concurrency control across diverse storage technologies while requiring only minimal memory overhead
Amorphous Nanomaterial: Emerging Energy Storage Materials
No full textThesis (Ph.D.)--University of Washington, 2025The accelerating impacts of climate change, combined with growing energy demands from artificial intelligence and the expansion of data centers, have created an urgent need for scalable and efficient battery technologies. Battery Energy Storage Systems (BESS) are critical for ensuring grid reliability, supporting renewable energy integration, and powering essential infrastructure. At the same time, the limited availability of lithium-ion batteries necessitates careful allocation toward applications that require both high energy density and power density. These include electric vehicles, medium- to heavy-duty trucks, delivery drones, and future aerial mobility vehicle such as eVTOLs. To meet the distinct needs of stationary and mobile storage, new material strategies are essential.This dissertation explores the use of amorphous nanomaterials to address key limitations in battery performance. Two case studies are presented. The first focuses on a non-stoichiometric sodium yttrium fluoride (NaYF-gel) developed as a fluoride-ion conducting electrolyte. The second investigates ultrasmall amorphous antimony sulfide nanoparticles as high-power-density negative electrode materials.
In the first study, we synthesized a nanoporous NaYF-gel by reacting NaF and YCl3 in water at room temperature. The resulting ion-rich product consists primarily of amorphous YF3, which exhibits short-range structural motifs similar to the cubic phase of NaYF4. This gel gradually incorporates Na⁺ ions and eventually crystallizes into the fully stoichiometric phase NaYF4. To halt this crystallization, we applied flash freezing followed by lyophilization, yielding a powder with high surface area that can be redispersed in organic solvents to form a gel with enhanced fluoride-ion mobility. Electrochemical tests demonstrated reversible cycling with low polarization when paired with calcium fluoride and metal electrodes, including copper, lithium, and sodium. Spectroscopic analysis revealed that solvent coordination and moisture content critically influence ion mobility, providing new insights into the transport behavior of this system.
In the second study, we developed a scalable room-temperature method to produce antimony sulfide nanoparticles smaller than 4 nm. These particles exhibited excellent performance in both Li-ion and Na-ion systems. The ligand environment during synthesis was found to influence both nanoparticle aggregation and electrochemical cycling behavior, despite being subsequently replaced with a different ligand. Replacing surface ligands or modifying coordination with molecules like trioctylphosphine altered its packing behavior and electrochemical performance. Compared to crystalline antimony, the amorphous form showed improved stability, which we attribute to reduced internal stress and possible thermodynamic factors. Pyrolysis of organic ligands created conductive carbon shells, further enhancing electrode conductivity.
Together, these studies demonstrate the versatility of amorphous nanomaterials in improving transport properties, mechanical stability, and improved manufacturability. These findings support the continued development of tailored nanomaterials for next-generation battery systems across both stationary and mobile energy storage applications
Development of an automated human whole-blood-based carbon nanotube paper-composite (CPC) hemostasis evaluation system
No full textThesis (Ph.D.)--University of Washington, 2025Hemostasis is a complex series of physiological reactions that occur when bleeding happens or a blood vessel is damaged. It functions to seal the injury site and stop bleeding. Hemostasis involves multiple interrelated steps, including platelet plug formation, coagulation, and fibrinolysis involving plenty of blood components, including platelets, coagulation factors, and erythrocytes (red blood cells). To date, no individual or model can fully explain all the reactions, mechanisms, and activities of the hemostatic process. Compared to the intricacies of hemostasis, effective hemostasis management is essential across a range of medical fields, including screening for cardiovascular diseases, monitoring anticoagulant therapy, blood transfusion management, preoperative hemostasis evaluation, Extracorporeal blood circuit therapy, screening for pre-existing coagulopathies. Therefore, a rapid, accurate and comprehensive hemostasis assessment is crucial for patient management and further hemostasis research. Currently, the most widely used routine hemostasis assays require separate tests for every hemostasis function. There are some novel global hemostasis assessments, including thromboelastography (TEG) and thromboelastometry (ROTEM). Limitations on recent assays, including partial or incomplete information on hemostasis function evaluation, high cost, time-consuming, labor-intensive, requiring professional medical personnel to conduct, analyze results. In this dissertation, an automated whole blood capacitance assessment using a carbon nanotube paper-composite(CPC) sensor is developed. Its automated module and integrated fluctuation analysis algorithms will also be presented.The automated whole-blood capacitance assessment system utilizes a carbon nanotube paper-composite (CPC) sensor for coagulation evaluation. Carbon nanotubes are emerging materials known for their high sensitivity, attributed to their excellent electrical conductivity and intricate surface structure, applied in many biosensors. In this study, carbon nanotube are used as the electrodes of a capacitance sensor to monitor changes in blood sample permittivity during the coagulation process the applied electric field. Importantly, the electrodes do not directly contact the blood samples. Compared to TEG, ROTEM, and other global or dielectric-based coagulation sensors, this system is low-cost, reusable, and portable. From the capacitance signal, multiple parameters can be analyzed and correlated with various hemostatic components, including coagulation function (factors), platelets, and erythrocytes(RBCs). The results also demonstrated strong agreement with the thromboelastography assay. Further investigations into the potential functions and advantages of this system are also presented in this dissertation.
The most commonly used routine hemostasis assays are typically limited to hospitals or clinical laboratories. These limitations arises primarily from two factors. First, the devices used in conventional hemostasis assays are non-portable and the testing process is complex. Routine tests require individual assays for each hemostatic parameter, with each test depending on specialized equipment. Also, these assays require trained medical personnel to perform the procedures and interpret the results, which often provide only partial information that must be analyzed by a medical professional. These two factors make the portability of conventional hemostasis assays nearly unfeasible. The development of this automated whole blood capacitance-based hemostasis evaluation system aims to address this issue by enabling portable hemostasis assessment. Automatic module can complete the entire process from blood sample aspiration to post-test disposal without manual intervention. Automatic assessment will reduce the requirements for medical personnel and helps standardize the testing protocol. The system's ability to perform multi-parameter and automated assessments makes hemostasis evaluation applicable in portable, cost-effective, and potentially home-use scenarios. Furthermore, this dissertation presents a novel fluctuation analysis of the capacitance signal. Fluctuation signal can avoid the influence of some noise on the capacitance baseline. Future improvements to the system will focus on further fluctuation analysis and more coagulation parameters correlations. A more compressed and industrially designed automatic module is also in the plan for the portable and automated objective
Evaluation of Cavitation Abrasive Surface Finishing as a Surface Treatment for Metal Produced Through Laser Powder Bed Fusion
Get PDFThesis (Master's)--University of Washington, 2025Laser powder bed fusion (L-PBF) has emerged as a highly viable method for manufacturing metal structural components across various industries. However, the inherently rough surfaces and complex morphologies of L-PBF components, particularly those with vertical, upskin, and downskin orientations, necessitate post processing treatments to improve surface finish and integrity. Additionally, heat treatments used to control microstructure and mechanical properties often produce a surface oxide layer that requires removal. In this investigation, cavitation abrasive surface finishing (CASF) was employed to improve the surface quality and remove the oxide layer of LPBF Ti6Al4V components, with specific attention to how build orientation, presence of alpha case, and line of sight effected the ability of CASF to improve the surface texture, introduce residual stress, and remove material. Results showed that CASF reduced the average surface roughness from approximately 5 to 20 μm in the as-built condition (depending on orientation) to as low as 4 μm. The process also imparted compressive residual stresses up to 600 MPa and was capable of removing the alpha case from direct line-of-sight surfaces. Despite these improvements, treatment uniformity varied with surface orientation. downskin surfaces, characterized by higher initial roughness and more extensive coverage of partially fused powder particles were the most challenging to treat. Even after CASF, these surfaces achieved significantly lower compressive residual stress, apparently due to shielding effects of particles that limit cavitation impact. Overall, CASF demonstrated strong potential as a non-chemical alternative for post processing LPBF titanium components, offering both surface smoothing and beneficial compressive stress. However, optimization of treatment parameters is needed to improve uniformity across orientations and to further assess the fatigue performance of treated surfaces
Polymer Modification via Vapor Phase Infiltration for Chemically and Mechanically Resilient Membranes
No full textThesis (Ph.D.)--University of Washington, 2025In recent years, membrane separations have been widely applied in wastewater treatment and gas separations due to their low energy requirements. Compared to ceramic membranes, which are highly stable but costly to manufacture, polymeric membranes have the greatest market share because of their low cost and scalable production. However, many of these polymers are unstable in organic solvents and at high temperatures, which has limited their use in applications requiring harsh conditions, such as organic solvent nanofiltration. An alternative class of membrane materials is inorganic-organic hybrid composite membranes, which exhibit enhanced stability compared to their additive-free polymers. In this work, we explored the modification of polymeric membranes with vapor phase infiltration (VPI), a post-synthetic modification technique that uses reactive vapors to enhance key polymer properties by incorporating inorganic or hybrid components. First, we studied the economic feasibility of membrane modification via VPI, estimating the cost of the VPI process based on similar techniques used in industry. We then investigated the influence of metal oxide infiltration on the chemical stability and mechanical properties of a common membrane support material, polyethersulfone, showing that infiltration depth and loading of composite strongly affect membrane stiffness and resilience. To mitigate the measured loss of ductility caused by metal oxide incorporation, alternative VPI chemistries such as metal alkoxides were explored, demonstrating successful integration of organic reactants into polymers. Finally, to accelerate development and optimization of new vapor phase processes, a high-throughput multiplexing reactor design was developed. Overall, these studies demonstrate the versatility of VPI as a cost-effective approach for tailoring polymeric membranes
Developing an Acquisition and Analytical Ecosystem for Protein Bioanalysis Using Mass Spectrometry and Programmed Temperature Control
Get PDFThesis (Master's)--University of Washington, 2025Temperature-controlled electrospray ionization (tcESI) coupled with mass spectrometry (MS), which allows for the measurement of mass-to-charge (m/z) ratios, enables the analysis of both structural information and stability in proteins, specifically by evaluating their thermal denaturation, a measure of protein unfolding. In this work, programmed-temperature electrospray ionization (ptESI) source, is used to continuously and rapidly heat and cool a nanoESI capillary containing proteins in solution. Combining MS and ptESI, enables the tracking of structural changes in unfolding and refolding events from solution-phase chemistry. In ion mobility (IM) measurements, charged ions under the influence of an electric field experience collisions that allow for separation based on size, shape, and charge, which yields additional insights into protein structure. During collision-induced unfolding (CIU), collisional activation allows IM-MS to probe gas-phase ion structures regarding their stability. Shifts in stability, when assessing collisional activation, may be reflective of structural changes. In chapter 1 of this work, the tools traditionally used for assessing protein structure and stability are described. In chapter 2, the acquisition software updates, supporting usability and maintenance, for the ptESI source are characterized. Next, in chapter 3, the programmatic analysis of the structural and stability information generated using the acquisition software is described. Then, in chapter 4, experiments utilizing ptESI, CIU, and ptESI coupled with CIU are characterized. Finally, in chapter 5, an experiment utilizing ptESI coupled with collision-induced unfolding is introduced, describing the application of the acquisition software and the subsequent programmatic analysis. Overall, this work marks a significant programmatic milestone towards automating the collection of both MS and IM-MS data utilizing the ptESI source. This work represents significant milestones towards being able to screen protein libraries or protein-ligand libraries in drug discovery and development.
Appendix A supplements Chapter 2 and contains a render of the physical ptESI source. Appendix B supplements Chapter 3 and contains examples of input and output data and visualizations related to the analytical pipeline. Appendix C supplements Chapter 4 and contains additional visualizations to assist with the characterization of ptESI, CIU, and ptESI coupled with CIU work
Hidden versatility, competition, and cooperation: Methanogenesis and nitrification drive unexpected carbon and nitrogen cycle linkages
No full textThesis (Ph.D.)--University of Washington, 2025Wetlands serve as crucial nodes in the carbon (C) and nitrogen (N) cycles by sequestering ~30% of global soil organic matter (SOM) and contributing to ~20% of annual global methane emissions despite accounting for only ~6% of global land area. The degradation, release, and retention of organic carbon and nitrogen in wetlands is facilitated by the activity of microorganisms that assemble into microbial communities within the pore network of wetland sediment aggregates. However, biogeochemical nutrient gradients will impose selective pressures that control the structure and function of the microbial community responsible for greenhouse gas emissions and SOM storage. The complex feedback between microbial activity and biogeochemistry makes it difficult to predict how increasing anthropogenic activity will influence nutrient cycling in wetland ecosystems. This dissertation aims to address this critical gap by linking single-cell microbial ecophysiology and community structure to ecosystem-scale biogeochemistry while developing predictive understanding of how microbial communities respond to environmental change. This work aims to investigate both the microbial interactions governing the degradation of organic carbon to methane by methanogenic communities in wetland sediments and the ecophysiology of ammonia-oxidizing microorganisms (AOM)—major contributors to terrestrial nitrous oxide (N2O) emissions and the global nitrogen cycle. Methane (CH4) and N2O are potent greenhouse gases (GHG) with global warming potentials 27 and 273 times greater than carbon dioxide (CO2) over a 100-year timescale, respectively. These important environmental processes were studied in laboratory enrichments and isolates with a combination of culture-based, isotopic, multi-omic, and mathematical modeling approaches. Wetland sediment chemostat enrichments were used in the first two sections to investigate spatial community structure, carbon degradation, and links between the carbon and nitrogen cycle. This thesis begins with a comprehensive overview of methanogenesis and nitrification and their connections to global carbon and nitrogen cycles within wetland ecosystems (Chapter 1). Synthetic wetland sediment aggregate chemostat enrichments and mathematical models of lactate degrading methanogenic communities demonstrated spatial community structuring and highlighted the importance of spatial proximity for methanogenesis (Chapter 2). Aerobic and anaerobic necromass degrading wetland communities reverted to their prior states after oxygen perturbations demonstrating the resiliency of these functionally redundant communities, while emphasizing necromass recycling as an underappreciated link between the carbon and nitrogen cycle (Chapter 3). Next, the physiological mechanisms underlying nitrogen removal in a novel partial nitritation-anammox reactor using high affinity AOM were evaluated (Chapter 4). This physiological investigation led to the discovery of hydroxylamine-dependent nitrate reduction and the biotic mechanism of N2O production by the complete ammonia-oxidizing (comammox) bacteria Nitrospira inopinata (Chapter 5). These results highlight the spatial dimension of metabolic niche partitioning, link the carbon and nitrogen cycles through necromass degradation, and uncover previously unknown nitrogen cycling metabolic versatility
The Virtue of Writing: Issues of Writing in the Literary Mind of Lu Ji's 陸機 (261-303 CE) "Wenfu" 文賦
No full textThesis (Master's)--University of Washington, 2025The "Wenfu" 文賦 is a fu-style writing composed by Lu Ji 陸機 (261-303 CE) around 300. Born into a prominent family of the State of Wu during the Three Kingdoms period, Lu Ji experienced the tragic loss of his home country at the age of twenty. After a decade of retreat, Lu Ji moved to the capital of the Western Jin dynasty and, in just a little more than ten years he had before his execution at the age of forty-three, produced numerous literary works, earning a place among the great literary figures of the Six Dynasties. The Wenxuan 文選, an anthology of poetry and prose from the Southern Liang Dynasty, preserves 61 of his works, and among these, the most famous is the "Wenfu," a treatise on literary creation which proclaims the principles of literary art in a highly ornate language. 3 This paper discusses the "Wenfu" in the context of literary criticism in Early Medieval China. After reviewing the biography of Lu Ji and the textual history of the "Wenfu" (Chapter 1), I will the trace the changes in the concepts of "wen" 文 and "writing" in Ancient China from the reign of Emperor Wu of the Western Han through the end of the Eastern Han, which culminate in the "Wenfu" (Chapter 2). Specifically, this chapter aims to elucidate the nature of the traditional Chinese literary historical discourse of the "self-consciousness" and the "independence" of literature in the Wei and Jin periods, a view that has been widely accepted since Lu Xun 魯迅. It also aims to identify the shift from "text" to "writing/literature" in Early China around the turn of the first century CE, collating recent scholarship. The purpose of this paper is to place the "Wenfu"—and not Cao Pi's "Lunwen"—as hallmark of the completion of this transition. For this purpose, this paper focuses on the concepts of "yongxin" 用心 and "xuanlan" 玄覧 in Chapter 3, discussing them through comparison with literary theories found in the Han dynasty's Huainanzi 淮南子 and Lunheng 論衡 and the so-called "xuanxue" 玄學 of the time. It will then discuss the issue of intertextuality and anxiety of writers in Early Medieval China (Chapter 4). In conclusion, this paper points out that the Ru-ist virtue is evident in the "Wenfu," and argues that the conventional view that literature became independent from Confucianism (morality) during the Wei and Jin periods requires some reservation (Chapter 5). The rhyme scheme and textual variations of the "Wenfu" are presented at the end (Appendixes)