Washington University Medical Center
Washington University St. Louis: Open ScholarshipNot a member yet
26344 research outputs found
Sort by
Robust 6D Fluorescence Microscopy
Get PDFSingle-Molecule Orientation Localization Microscopy (SMOLM) measures the positions and orientations of single fluorophores precisely, approaching fundamental classical and quantum limits. However, a disadvantage of SMOLM is the time-consuming nature of its acquisition and reconstruction processes. In my thesis, I introduce an innovative framework that simplifies the detection and estimation algorithms used in SMOLM into a computationally efficient high-dimensional deconvolution algorithm. This framework extracts six-dimensional information of continuous biological structures from just a single camera image (i.e., a single shot). While this method is diffraction limited, unlike conventional SMOLM algorithms, it nevertheless offers accurate and precise estimations of the orientations of collections of molecules. Crucially, it is suitable for capturing dynamic changes in biological structures, thereby broadening its applicability in scientific investigations
Reflections, Accomplishments, and Opportunities: The CSD Annual Letter
Get PDFIn this annual letter to mark the start of the Center\u27s for Social Development\u27s 30th year, Founding Director Michael Sherraden shares reflections on some recent accomplishments and discusses the opportunities ahead
Learning-based Artifacts Removal Priors for Computational Imaging
Get PDFThis dissertation addresses the integration of physical models with learning-based artifact removal priors for computational imaging. The motivation for our work stems from the growing interest in combining imaging models, which ensure data consistency with observed measurements, with deep learning, which provides advanced data-driven prior modeling. Following this approach, we adopt classic statistical inference as the foundational framework and integrate deep artifact removal models as image priors. This allows our framework to simultaneously leverage both physical models and learning-based priors. Furthermore, the increasing size of images and measurements in modern computational imaging systems imposes significant computational and memory burdens. Another objective of this dissertation is to extend our framework to these scenarios by incorporating large-scale optimization techniques. We have developed multiple algorithms to achieve efficient and reliable imaging. We validate the performance of our algorithms through rigorous theoretical analysis and by demonstrating their effectiveness in various imaging applications. The dissertation significantly extends three algorithmic frameworks—that is, plug-and-play priors (PnP, Part II), deep model-based architectures (DMBA, Part III), and diffusion probabilistic models (DPM, Part IV)—with multiple contributions including the design of novel algorithms, establishment of unified theory, and applications to real imaging problems. Specifically, in Part II, we present in-depth discussions of two popular PnP algorithms: the proximal gradient method PnP (PnP-PGM) and Regularization by Denoising (RED). Our contributions include the introduction of Regularization by Artifact Removal (RARE), which broadens the current denoiser-centric view of PnP methods by considering priors corresponding to networks trained for more general artifact removal. We provide recovery analysis under both deep denoising priors and artifact removal priors for compressive sensing scenarios. Additionally, we propose scalable PnP variants that efficiently infer large images using block coordinate computing techniques. Finally, we introduce a new method called Calibrated RED (Cal-RED), which enables the joint calibration of the measurement operator along with the reconstruction of the unknown image. In Part III, we conduct both empirical and theoretical investigations on DMBA. Building on existing variants of DMBA—Deep Unfolding (DU) and Deep Equilibrium Model (DEQ)—which interpret the iterations of a model-based algorithm as layers of a deep neural network and train it end-to-end in a supervised fashion, we propose scalable DMBA variants for processing a large set of measurements using online gradients. Our analysis framework, based on monotone operator theory and stochastic gradient descent in machine learning, is unified for online DMBA and represents a novel contribution to the existing literature. Finally, we develop a new DMBA for learning explicit regularization functionals, extending the existing implicit prior-centric approaches of PnP and DMBA. In Part IV, we extend DPM—a novel probabilistic prior—to computational imaging by developing a conditional diffusion posterior sampling method. Our method, DOLCE, explores the regularization capability of conditional DPM specifically for the extremely ill-posed inverse problem of limited-angle computed tomography (LACT)
Robust quantitative photoacoustic imaging for colorectal cancer treatment monitoring
No full textColorectal cancer is the 2nd leading cause of cancer death in the United States and the incidence among population under 50 years old has been increasing for the past decade. Contrary to obvious vascular and morphological structures in pre-treatment colorectal cancer, during treatment and post-treatment colorectal cancer tissue often contains fibrosis and edema in tumor bed. These complications cause difficulties in evaluating tissue response to neoadjuvant chemoradiation therapy with MRI and endoscopy, leading to unnecessary resection of rectum, which increases healthcare cost and impacts patient quality of life. Photoacoustic (PA) imaging is a promising tool for observing blood distribution in tissue without the need for exogenous contrast agents. It has been shown to have potential in colorectal cancer treatment monitoring, but prototype systems need to be optimized for reliable patient studies. To address this challenge, development of a robust PA endoscopic imaging system is needed, as well as extraction of quantitative functional information based on signal processing pipelines. This thesis work first improves upon the acoustic resolution photoacoustic and ultrasound imaging system developed previously, with respect to system electronics, mechanics, programming, and optics. The system’s maneuverability is greatly enhanced, and signal integrity is significantly improved. Then a second-generation system was developed for higher portability and reduced footprint, both systems were characterized thoroughly for its performance, and the processing pipeline was optimized with signal from tissue mimicking imaging phantoms. Two quantitative functional information processing pipeline was implemented. The former is a photoacoustic doppler processing to obtain blood flow information from in vivo patients. The algorithm is validated with flow phantoms and used to study blood flow profile in vivo. The latter is a photoacoustic and ultrasound elastography processing algorithm. Tissue mimicking elastography phantoms were developed to assess the algorithm, which is latter applied to ex vivo imaging of colorectal tissue. Both the PA doppler and PA & ultrasound elastography offer added quantitative functional information to morphology and tissue vascular distribution. The design and implementation of the PA endoscopic imaging system as well as photoacoustic doppler and photoacoustic elastography represent readiness in clinical translation of photoacoustic imaging for rectal cancer treatment monitoring. From experimental characterization to validation and optimization via ex vivo imaging, to in vivo imaging in operating room and endoscopic suites in the future, the PA endoscopic imaging system has shown its potential in clinical application and suitability to future large scale clinical trials
Modeling the Performance and Resource Requirements for Gamma-Ray Telescope Signal Processing
Get PDFThis thesis investigates the buffering requirements of the data pipeline for the Advanced Particle-astrophysics Telescope gamma-ray telescope. Given the importance and stochastic arrivals of astronomical signals, it is crucial to ensure that each gamma-ray signal is fully buffered to prevent data loss. To achieve this, FIFOs are inserted into the data processing pipeline to prevent bottlenecks during data packet processing. Buffers play a critical role in regulating data flow, preventing data loss, and ensuring efficient data processing. They act as temporary storage areas, absorbing data surges and releasing it steadily, thus maintaining the pipeline’s optimal performance. In general, buffer integration significantly enhances the stability and reliability of a data processing system. The use of buffers ensures the integrity and timeliness of critical astronomical signals, pro- viding reliable data support for subsequent investigation. This thesis uses discrete-event simulation models to assess the buffering requirements for a prototype gamma-ray telescope computational pipeline, ADAPT, the Antarctic Demonstrator for the Advanced Particle- astrophysics Telescope
Impacts of Phosphate and Residual Free Chlorine on the Precipitation, Stability and Dissolution of Lead Corrosion Products
Get PDFLead release from lead pipes into drinking water is a critical threat to public health. Orthophosphate addition is an established method of controlling lead concentrations in tap water. Many phosphate-based inhibitors currently used by utilities are blends of orthophosphate and polyphosphate. Orthophosphate can limit lead release from pipes by forming low-solubility lead phosphate solids on pipe inner surfaces. In contrast, polyphosphate can accelerate lead release by forming soluble lead-phosphate complexes. Due to the potential conversion from polyphosphate to orthophosphate in the distribution system and to complex dynamics of water chemistry in lead service lines (LSLs), the net impact of a blended phosphate chemical on lead in tap water had been challenging to assess. In addition to lead-phosphate solids, lead(IV) oxide (PbO2) solids in the scales of LSLs can play an important role in controlling lead concentrations. PbO2 has an extremely low solubility (below 0.1μg/L for most situations) that has the potential to maintain desired low lead concentrations in water. However, PbO2 is only formed and stable at a high redox potential environment, as Pb(IV) is highly oxidative and can easily undergo reductive dissolution. Free chlorine, as an important tap water infectant, is a strong oxidant that can help form and maintain PbO2 in LSLs. However, the threshold free chlorine to maintain PbO2 had been determined for layers of PbO2 on the surfaces of elemental lead materials. There were also unresolved questions regarding the impacts of stagnation time and orthophosphate dosing on the stability of PbO2 in scales of harvested lead pipes. This dissertation includes five studies aimed at providing a better understanding of the role of phosphate and residual free chlorine on the precipitation, stability and dissolution of lead(IV) oxide and lead-phosphate solids at conditions relevant to LSLs involved in residential drinking water supply. In the first study, the effects of polyphosphate presence as a component of a mixture with orthophosphate on lead release were investigated on harvested lead pipes. This study confirmed that orthophosphate will be a better corrosion control choice for LSLs than a blended phosphate for a system that has a major goal of lead corrosion control and no specific need for metal sequestration. The second study investigated the stability of PbO2 on the surface of metallic lead coupons and its potential role in corrosion control. This study found that for systems that have PbO2 already present in lead scales, promoting conditions that maintain PbO2 stability could be an effective method for limiting lead release to drinking water. A threshold concentration of residual free chlorine was identified to maintain PbO2 stability. In the third study, the effect of residual free chlorine and water stagnation on lead release associated with PbO2 dissolution was evaluated with actual LSLs harvested from a public water system. Similarly to in the study with lead coupons, this study with lead pipes identified a threshold free chlorine concentration below which PbO2 undergoes reductive dissolution. Long water stagnation times (five days) can exacerbate the lead release from pipes into water especially for particulate lead. Scale analysis reveals that plattnerite (β-PbO2) was the primary component in the pipe scales and that most PbO2 pipe scales maintained their integrity during periods of water stagnation even when stagnation did result in substantial increases in dissolved lead. The fourth study investigated the effect of orthophosphate on lead release from PbO2-rich layers on lead pipes and coupons. The study found that the use of orthophosphate in LSLs with rich PbO2 layers in scales may not provide a substantial benefit for lead corrosion control during the initial phase of its addition. The fifth study explored the variations in lead pipe scales from 43 different drinking water systems based on laboratory scale analysis together with data from previous research. It also evaluated the applications and limitations of chemical equilibrium models for predicting the solid phases present in lead pipe scales. The study found that the compositions of lead pipe scales depend on water chemistry, corrosion control methods, and disinfectant types. While equilibrium predictions are not completely predictive and have certain limitations, they remain valuable for making initial assessments and screening lead solids in pipe scales
Human-Environmental Dynamics in prehistoric Southwest China: Environmental Change and Chronology at Haimenkou, Yunnan
No full textThis dissertation investigates the environmental dynamics and chronological developments at Haimenkou, Southwest China, providing a nuanced understanding of human-environment interactions from the late Pleistocene through the Holocene. Utilizing geoarchaeological surveys, particle size analysis, loss-on-ignition, and diatom analysis, this research delineates shifts in landscape and settlement patterns in relation to climatic changes and human activities. The results indicate that adaptation to the wetland environment was gradual, with the extent of landscape modification intensifying over time. This modification involved forest clearance, leading to severe soil erosion on mountain slopes, which subsequently destroyed and buried neighboring settlements to Haimenkou. Through rigorous stratigraphic and radiocarbon analysis, the dissertation refines the chronological framework of settlement and abandonment at Haimenkou, correlating these with broader settlement patterns across mainland Southeast Asia. A novel statistical tool is applied to test for chronological gaps, confirming that the observed gap was due to an occupational hiatus rather than insufficient data or biased sampling. The study addresses a critical phase known as the “Missing Millennia”, a period before the occupation of Haimenkou marked by scant archaeological records and significant ecological and geomorphological transformations in the Mid-Holocene. The findings suggest that the Haimenkou site, positioned strategically along the river systems at the eastern edge of the Tibetan Plateau, served as a dynamic locus for early human settlement, agricultural development, and cultural transitions in mainland Southeast Asia. These results underscore the complex interplay between climatic forces and early human landscape management in shaping the archaeological preservation and environmental history of Southwest China, contributing valuable insights into the broader prehistoric dynamics of Southeast Asia
Transcriptional precision in photoreceptor development and diseases - Lessons from the retinal transcription factor CRX
No full textPrecise gene expression programs are foundational to organism development and homeostasis. Transcription factors play a central role in establishing and maintaining these programs. Large-scale transcriptomic studies have generated parts lists of transcription factors. However, these descriptive works rarely formulate a cohesive model of how these transcription factors work together to support a functional organism. Mutations that affect transcription factor functions can lead to a broad range of human diseases. Therefore, it is crucial to understand the principles that transcription factors achieve spatial, temporal, and quantitative precision in their regulated gene expression programs. I have investigated transcriptional regulation in rod and cone photoreceptors. Photoreceptors are specialized neurons in the retina that function in the initial step of vision. Development and maintenance of photoreceptors are regulated by a gene regulatory network that revolves around a homeodomain transcription factor (HD TF) encoded by the Cone-rod homeobox (CRX) gene. CRX regulates photoreceptor gene expression by recognizing specific HD DNA motifs in the genome. The conventional model of the CRX function posits that it binds one type of consensus DNA motif. However, human mutations in CRX HD are associated with distinct blinding diseases, suggesting complex modes of CRX HD-DNA interactions and differential perturbations by disease-associated mutations. My thesis project aims to establish quantitative models to understand the features of CRX HD-DNA interactions in photoreceptor development and diseases. I have focused my research on two CRX HD missense mutations, p.E80A (E80A) and p.K88N (K88N), that are associated with dominant Cone-rod dystrophy (CoRD) and dominant Leber congenital amaurosis (LCA) in humans. Using an integrated approach that combines quantitative in vitro biochemical models, functional genomics, cellular profiling, and functional testing in mutation knock-in mouse models, I identified two novel gain-of-function pathogenic mechanisms. I found that E80A and K88N differentially alter CRX HD-DNA binding specificity and cooperativity, leading to distinct photoreceptor developmental deficits in mutant mouse retinas recapitulating human conditions. By comparing the epigenome and transcriptome dynamics in WT and Crx mutant retinas, I uncovered an underappreciated role of differential CRX interactions with subtypes of HD motifs in regulating stage and cell-type specific chromatin remodeling and temporal gene regulation during photoreceptor development. Collectively, my study refines the CRX mechanistic model in photoreceptor development, expands our knowledge of the diverse mechanisms that CRX mutations lead to severe, dominant retinopathies, and lays the foundation for the future development of therapeutic strategies targeting different pathogenic mechanisms. Beyond CRX, my study affords insights into the intricate mechanisms that transcription factors deploy to achieve precision and functional specificity in regulating diverse biology processes. Moreover, the holistic approach in my study offers a transferable framework for the mechanistic studies of transcription factors in other cell types and their associated diseases
Island Hopping in the Tropics: Conservation Genetics, Speciation, and Biogeography in the Iconic Flowering Plant Group Hibiscus sect. Lilibiscus
Get PDFIslands have played a crucial role in evolutionary biology, providing key insights for both Charles Darwin and Alfred Russel Wallace in formulating their theories of evolution. The development of Island Biogeography Theory and the General Dynamic Model of oceanic island biogeography have formed the theoretical foundation of island studies over the last sixty years, leading to substantial advancements in our understanding of biodiversity evolution. Despite their critical value for evolutionary theory, islands face unprecedented ecological collapse, with species on islands facing significantly higher extinction rates than on continents. The island extinction crisis highlights the need for integrating conservation perspectives into evolutionary research. This dissertation uses Hibiscus section Lilibiscus as a model group in which to examine three fundamental evolutionary processes — dispersal, speciation, and extinction — that have produced the incredible but vulnerable biodiversity exhibited by oceanic islands across the globe. Chapter 1 assesses ongoing conservation activities for the critically endangered Hibiscus liliiflorus in the Mascarene Islands, revealing a gap in the conservation of extant genetic diversity and making recommendations for improved conservation actions. Chapters 2 and 3 focus on speciation within the Mascarene Islands, resolving taxonomic confusion and producing a population-level genetic study that shows unique support for speciation predictions fundamental to the General Dynamic Model. Chapter 4 presents a global phylogeny of sect. Lilibiscus, revealing dispersal patterns that suggest islands may not be evolutionary dead ends, as is generally assumed, and may instead act as the launching point for the global diversification of the group. Together, these studies contribute to a deeper understanding of island biogeography while making a significant contribution to the conservation of sect. Lilibiscus. By linking evolutionary theory with practical conservation strategies, this dissertation underscores the importance of a holistic approach to studying and preserving island biodiversity
Investigation of PHOTOPERIODIC CONTROL OF HYPOCOTYL 1’s Role in Promoting Phytochrome B-Photobody Phase Separation
No full textLight is arguably the most important environmental stimuli for all of plant growth and survival. Not only does light serve as an energy source via photosynthesis, it also serves as a signal for developmental programming. Plants perceive light through light-responsive proteins termed photoreceptors. Phytochromes are conserved red/far-red light photoreceptors. In the model plant Arabidopsis thaliana (thale cress), phytochrome B (phyB) is the primary red light photoreceptor. PhyB is involved in regulating almost all developmental processes throughout the plant’s life cycle, from seedling germination to flowering. In response to red light, phytochromes photoconvert to a biologically active form, transport from the cytoplasm into the nucleus, and further compartmentalizes to form biomolecular condensates termed photobodies. Accumulating evidence indicates that phyB photobodies are essential for full-strength phytochrome-mediated signaling and physiological outputs. The Nusinow lab identified a protein, PHOTOPERIODIC CONTROL OF HYPOCOTYL 1 (PCH1), that is sufficient for stabilizing phytochrome B in its active conformation in vitro and necessary for photobody formation in vivo. However, how PCH1 promotes the phase separation of phyB into photobodies remained unknown. Here, using multiple high-resolution microscopy techniques, I show that phyB photobodies are liquid-liquid phase separated (LLPS) biomolecular condensates, and that PCH1 alters photobody dynamics and phyB molecule partitioning from the nucleoplasm into photobodies. I hypothesized that PCH1, through its predicted disordered protein structure, drives phyB photobody phase separation. Combining microscopy, genetics, and biochemistry, I characterized PCH1 protein domains. My results show that PCH1 is sufficient to phase separate, and that PCH1’s C-terminal regions are important for promoting PCH1-phyB nuclear localization and phase separation. My results show that PCH1 oligomerizes in vitro and suggests that phyB preferentially binds to PCH1’s predicted intrinsically disordered regions (IDRs) containing helices. Overall, I propose the following model in which PCH1 is a disordered protein that promotes phyB’s nuclear localization and modulates its phase separation to control Arabidopsis hypocotyl growth. Lastly, I helped develop an Expansion Microscopy method in Arabidopsis protoplasts that can be used for studying biomolecular condensates. Understanding how plants perceive light at the molecular level will help to identify genetic targets for engineering maximized sunlight capture of agriculturally important crops