22689 research outputs found
Sort by
THE COMMITTEE ON FOREIGN INVESTMENT IN THE UNITED STATES: THE EVOLUTION OF AN INSTITUTION
No full textThe Committee on Foreign Investment is the United States (CFIUS), the interagency body charged with evaluating all incoming foreign direct investment (FDI) for national security risk, is an institution that has evolved significantly from its beginnings. It began as an advisory body in the mid-1970s but is now an opaque institution with considerable influence. This study intends to examine the periods during which the laws and regulations regarding foreign investment strengthened, providing CFIUS with enhanced statutory and regulatory power. This will be done using critical juncture analysis methodology. By examining the periods just preceding and during these changes, this study will attempt to identify the factors that made these periods result in legislative or regulatory change. In applying critical juncture analysis, these factors will be viewed through the lens of critical antecedents, permissive conditions, and productive conditions.
The goal of this study is to 1) provide a historical survey of CFIUS, 2) analyze the critical periods of regulatory change and their characteristics, and 3) identify commonalities between these periods
Cellular Responses to Topological and Electrical Stimuli: Mechanisms of Migration and Division
No full textCollective cell movement and organization in response to external cues are essential for wound healing, tissue development, and regeneration. In this thesis, I use computational techniques to investigate the collective behavior of cells in response to topographical and electrical cues, focusing on elongated cells that align with their neighbors to create local nematic order.
Recent experiments with cell monolayers have highlighted the importance of areas where nematic order is disrupted, known as topological defects. These defects, characterized by their charge in 2-D, influence cell apoptosis and morphogenesis of tissues. I investigate fibroblast organization, motion, and proliferation on substrates with micron-sized topographical patterns that induce topological defects with charges of +1 and -1 using simulations. Unlike earlier studies on other cell types, our findings show that the increased density of fibroblasts at +1 defects and decreased density at -1 defects are not due to collective migration. Instead, different division rates based on cell area and aspect ratio lead to these density variations. Modeling cells as self-propelled deformable ellipses interacting via a Gay–Berne potential, our simulations capture key experimental features: high cell density at +1 defects, low density at -1 defects, deterioration in alignment at higher densitis, and varied morphologies near defects.
I also examine the effects of cell anisotropy and nematic order when elongated cells are exposed to an external electric field. Both elongated and non-elongated cells tend to migrate parallel to the electric field, a process known as galvanotaxis, which occurs naturally during wound healing and embryogenesis. Motivated by theoretical studies on elongated cells sensing chemical gradients, I model cells to have varying accuracy in sensing the electric field based on their orientation. Given that cells tend to align their long axes, I explore whether this alignment helps them sense the electric field more accurately. Simulations show that if cells orient perpendicular to their average velocity, aligning their long axes with nearest neighbors enhances directional response to electric fields. However, this benefit requires that sensing accuracy be highly dependent on cell orientation. My findings also indicate that cell-cell adhesion modulates the accuracy of the group's response to the electric field
Towards precise, functional, and non-contact clinical intervention through ultrasound and photoacoustic approaches
No full textUltrasound is widely utilized in various clinical fields, including urology, OB/Gyn, cardiology, and breast imaging, owing to its real-time imaging capabilities. Its ability to provide real-time tissue morphology has made it indispensable in surgical settings. Additionally, photoacoustic imaging has emerged as a promising modality, offering optical contrast by exploiting the distinct optical properties of different tissue types.
Beyond imaging, ultrasound also plays a therapeutic role by leveraging the mechanosensitive activation of ion channels within tissues through acoustic waves. The interaction between high acoustic pressure and tissue can lead to neuronal activation or modulation, offering potential treatments for diseases such as Parkinson’s.
Despite its diverse clinical applications, ultrasound faces a fundamental limitation: the need for the transducer to be in direct contact with the target (e.g., skin) to minimize acoustic impedance mismatch between air and tissue. This requirement restricts the broader application of ultrasound.
This dissertation investigates various non-contact applications of acoustic waves. Non-contact photoacoustic stimulation was developed to restore visual function in retinal disease treatment. Furthermore, limitations in laparoscopic prostatectomy were addressed by integrating high-resolution ultrasound imaging with functional nerve morphology via photoacoustic voltage-sensitive dye imaging. A photoacoustic marker-based frame registration between the endoscopic camera and ultrasound imaging enabled automatic tracking of target imaging slices during surgery. The same photoacoustic source was utilized for microscopic photoacoustic imaging. Lastly, a fully non-contact ultrasound imaging system was developed by harnessing the photoacoustic effect and laser interferometric technology, enabling full non-contact imaging of tissue morphology, including the first-ever introduction of blood flow estimation in non-contact laser ultrasound
Self-Supervised and Weakly-Supervised Computer Vision Methods for Lensless Imaging
Get PDFComputer vision (CV) enables systems to understand visual information, which is key for medical applications such as image-assisted interventions and image-based diagnosis. Point-of-care testing (POCT) through lensless imaging is one promising such application, which allows rapid and cost-effective assessment of clinically relevant variables by avoiding logistic delays associated with laboratory tests (e.g., complete blood count or urinalysis). As POCT devices image biospecimens, CV algorithms extract relevant information from the images by, for example, performing cell detection and classification. Recent advances in CV have been largely mediated by the availability of large, annotated datasets. However, annotating medical data can become prohibitively expensive as it requires highly trained experts. Moreover, suitable experts might not exist for innovative imaging modalities like lensless imaging. Therefore, there is a need to develop self-supervised or weakly-supervised CV methods that can operate without detailed annotations. Besides its use in POCT, lensless imaging is an interesting case study as it poses additional challenges to CV including low resolution, potentially large number of objects per image, and the need for image reconstruction.
This dissertation focuses on developing self-supervised and weakly-supervised CV methods for lensless imaging, exploring the hypothesis that embedding domain knowledge into learnable models can alleviate the need for detailed annotations. In particular, the contributions of this thesis are the following. First, we propose Adaptive Sparse Reconstruction (ASR), the first self-supervised approach to jointly perform phase retrieval, reconstruction, and point spread function estimation, improving the robustness of reconstruction methods to imprecise knowledge about the imaging system. Second, we propose Deep Sparse Detector (DSD), a self-supervised cell detector that utilizes data priors and physics-based forward models to perform cell detection in thin and thick specimens. Finally, we leverage label proportions as a weak form of annotation (at the subject level) for cell classification, and propose an optimal transport-based loss, called Vertex Proportion (VP), to map subject-level class proportions to cell-level labels during training. We successfully validate our methods on real datasets and create synthetic datasets for detailed quantitative evaluation. Overall, these methods aim to boost the impact of POCT lensless technology, while also shedding light on other applications in which annotations are scarce but domain knowledge is abundant
Exploring ADP-ribosylation Forms Using Mass Spectrometry-Based Proteomics Methods
No full textADP-ribosylation is a post-translational modification (PTM) that involves the addition of one or more units of adenosine diphosphate (ADP)-ribose to a target protein, often occurring during the cellular stress response to a chemical toxin or viral infection. There are two main layers that facilitate ADP-ribosylation’s role in coordinating cellular processes. First, ADP-ribosylation can be added to 9 different amino acids of very different chemical properties. Second, ADP-ribosylation can exist in a monomeric (MARylation) or polymeric (PARylation) form. In addition, polymers of ADP-ribose (PAR) can have linear or branched structures. The change in site and form of ADP-ribosylation during oxidative stress has become much better characterized over the past decade. However, the lack of high throughput tools to monitor both the site and form of ADP-ribosylation within one experiment prevents deeper insights into other biological contexts in which this regulation is important. This thesis focuses on exploring ADP-ribosylation forms using mass spectrometry-based proteomics methods. The first chapter further explains the layers to regulation of ADP-ribosylation and compares and contrasts other PTMs that share similar challenges to mass spectrometry analysis as ADP-ribosylation. The second chapter details a proof-of-principle for a proteomics workflow to generate two mass tags to distinguish between MARylation and PARylation sites. The third chapter examines the ADP-ribosylome during influenza A infection and how a viral protein suppresses the polymeric form of ADP-ribosylation. The last chapter discusses alternative proteomics approaches inspired by other PTMs that have potential to successfully characterize the site and form of ADP-ribosylation
Investigating function and regulation of chromatin factors in the maintenance of germline identity
No full textProper germline proliferation and maintenance of genome integrity are critical, tightly regulated processes that ensure the faithful transmission of genetic and epigenetic information across generations. These processes rely on complex networks of protein machinery that coordinate gene expression in a cell type- and developmental stage-specific manner, largely through chromatin architecture. This coordination is particularly essential in the germline, where failure to establish or interpret appropriate chromatin states can lead to the loss of germ cell fate identity and sterility. In this work, we investigate how the heterochromatin-euchromatin landscape is communicated between chromatin-associated factors in the maternal germline of C. elegans.
The H3K36me3 reader MRG-1 is an essential protein whose loss leads to gene misregulation, germ cell fate loss, and ultimately sterility. We have found that MRG-1 binds along the gene bodies of actively expressed genes in the adult C. elegans gonad, mirroring in part the distribution of the euchromatin-associated histone mark H3K36me3. Additionally, MRG-1 is enriched at the transcriptional start sites (TSSs) of these genes. We hypothesize that MRG-1 is recruited to germline-active gene bodies through its canonical binding to H3K36me3, while its presence at TSSs may facilitate the recruitment or stabilization of transcriptional activating complexes that ensure proper gene expression. Consistent with this, we show that MRG-1 interacts strongly with two histone acetyltransferase complexes and the Set1/COMPASS H3K4 methyltransferase complex in the adult germline. The H3K36me3 and H3K4me3 histone modifications are known to be critical regulators of gene expression. Our data suggest that MRG-1, in coordination with transcriptional modifying complexes, represents a key mechanism of epigenetic communication between the maternal germline and the developing embryo.
Furthermore, we identified that the MORC-1, a chromatin-compacting factor that binds and regulates germline-expressed genes, is post-translationally modified by phosphorylation within an intrinsically disordered region by protein kinase CK2. CK2 is required for proper MORC-1 nuclear localization and may also modulate the liquid-like properties of MORC-1. Together, our studies provide new insights into the molecular mechanisms by which the chromatin factors MRG-1 and MORC-1 are regulated, ultimately helping to maintain the heterochromatin-euchromatin landscape of the germline genome
SPATIOTEMPORALLY RESOLVED IMAGING AND SPECTROSCOPY OF ZEOLITES AND ZEOLITIC IMIDAZOLATE FRAMEWORKS: DEPOSITION, DISSOLUTION, AND IRRADIATION-INDUCED MODIFICATIONS FOR THIN FILM MEMBRANES AND LITHOGRAPHY RESISTS
No full textThe solvent- and electron-beam- (e-beam-) sensitivity of Zeolitic Imidazolate Frameworks (ZIFs) can be detriments for their potential uses as adsorbents, catalysts, and photoresists and for the determination of their structure using electron microscopy and diffraction. However water- soluble ZIFs upon e-beam exposure are rendered insusceptible to facile dissolution by water, enabling their use as resists for electron-based lithography. Here, we use spatially resolved spectroscopy methods to determine spectroscopic signatures of e-beam induced changes in a prototypical ZIF-L. We confirm an earlier suggested two-stage dose-dependent evolution consisting of an amorphization stage at low doses, followed by a chemical altering stage at higher doses, and indicate that the former is characterized by hydrogen bond breaking, while in the later, high dose stage, e-beam induced 2-methylImidazole (2mIm) ring opening and dehydrogenation lead to the formation of a zinc isocyanide amorphous framework. The structural changes caused by e-beam treatment have consequences on the choice of solvent for pattern development. Notably, while the dissolution of ZIF-L in water slows down at high pH (achieved by addition of tetramethylammonium hydroxide (TMAOH) to the water solvent), ZIF-L treated at the low-dose regime (between 2-3 mC/cm2) undergoes accelerated dissolution in these basic solutions allowing the realization of positive tone resist behavior (i.e., dissolution of the e-beam treated areas and retention of the crystalline untreated ZIF-L). The generality of this dual-tone behavior seen with ZIF-L, i.e., the positive tone mode at lower doses when using TMAOH aqueous developer reported here, and the earlier reported negative tone behavior, is demonstrated for an amorphous ZIF (aZIF) film. Positive tone behavior of amorphous ZIF films in an aqueous solution is thus established for the first time. Furthermore, EUV was successfully used with aZIF for selective development
BONE-INSPIRED SELF-ADAPTIVE MATERIALS BY COUPLING STRESS WITH MATERIAL SYNTHESIS USING PIEZOELECTRIC EFFECTS
No full textMaterials naturally experience progressive degradation over time, resulting in diminished performance, reduced lifespans, and increased costs related to inspection, maintenance, and downtime. Additionally, various applications require unique combinations of properties, such as high load-bearing capacity and enhanced energy dissipation. However, conventional materials often encounter trade-offs, where enhancing one property compromises another, limiting the development of optimally balanced materials.
This thesis introduces liquid-infused porous piezoelectric scaffolds (LIPPS) as an innovative material inspired by bone remodeling, utilizing piezoelectric charge-induced mineralization to improve load-bearing capacity and energy dissipation. LIPPS address mechanical trade-offs, providing tunable stiffness, dissipation, and fatigue resistance.
A key advancement in LIPPS is the application of piezoelectric charge-induced mineralization to simultaneously enhance stiffness and energy dissipation under cyclic loading. For instance, after 12 million loading cycles, LIPPS achieve a 3,600% increase in elastic modulus and a 3,000% increase in hysteresis, overcoming conventional trade-offs. Additionally, LIPPS exhibit a reprogrammable stiffness distribution based on the spatial distribution of mechanical loading, enabling controlled self-folding governed by the amplitude and location of applied forces.
Significantly, LIPPS leverage a stress-coupled mineralization mechanism, where mineralization preferentially accumulates in high-stress regions, such as crack tips, reducing damage and slowing crack propagation by 90%. This mechanism substantially enhances fatigue resistance, with LIPPS demonstrating a 947% increase in fatigue threshold. Furthermore, LIPPS are not restricted to a single material system and are compatible with diverse platforms, including PDMS and hydrogels, expanding their potential applications.
Beyond polymer systems, this work explores metal composites through piezo-driven metal-based deposition, utilizing mechanically induced potentials to facilitate metal ion reduction. This approach enables remarkable electrical self-healing and mechanical reinforcement, offering new possibilities for self-adaptive conductive and protective coatings.
With extensive applicability, LIPPS represents a transformative material platform for soft robotics, biomedical engineering, structural materials, and next-generation smart materials. By addressing critical challenges in material performance and sustainability, LIPPS offers a pathway toward more resilient and adaptive engineering systems
UNDERSTANDING MOTOR NEURON BIOLOGY TO ADVANCE ALS THERAPEUTICS
No full textAmyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease primarily affecting motor neurons. Although most cases are sporadic (sALS), familial ALS (fALS) is linked to specific gene mutations. In this dissertation, I examine two mouse models to better understand ALS pathogenesis and therapeutic potential: the SOD1G93A model, carrying the first discovered ALS-linked mutation, and a TDP-43 conditional knockout (TDP-43cKO) model, which mimics the loss of TDP-43 function observed in 97% of ALS cases.
In chapter 2, we investigate the effects of a combinatorial therapeutic strategy to promote neuronal regeneration and prevent cell death in the SOD1G93A mouse model by modulating the dual leucine zipper kinase (DLK) signaling cascade. DLK, a master sensor of axonal injury, regulates both neuronal regeneration and cell death. We hypothesized that DLK deletion combined with overexpression of its proregenerative target, Atf3, would improve survival and motor outcomes. While this approach did not enhance survival, it preserved motor neurons and motor function until much later disease time points, suggesting that such combinatorial therapies may improve quality of life in ALS patients.
Next, in chapter 3, we focus on the consequences of TDP-43 deletion in motor neurons. TDP-43, a critical RNA binding protein, is mislocalized in a majority of ALS cases, disrupting RNA splicing and causing motor neuron dysfunction. Using the TDP-43cKO mouse model, we explore the early functional and transcriptomic changes following TDP-43 loss. We find activation of stress response pathways, dysregulation of cytoskeletal genes, and nuclear retention of mis-spliced transcripts. Surprisingly, motor neuron death occurs later in the disease course, providing a potential window for therapeutic intervention. The focus of chapter 3 was characterization of this mouse model to build a solid foundation for future translational studies. In appendix 1, I build on this work and describe a collaboration in which we use this mouse model to test therapeutic strategies specifically targeting TDP-43 deficient cells.
DEVELOPMENT OF METHODS FOR THE EVALUATION OF CELL-FREE DNA IN INDIVIDUALS WITH CANCER, AUTOIMMUNE, OR VASCULAR DISEASES
Get PDFArtificial Intelligence (AI) is now a cornerstone of modern dataset analysis. However, quantification of the confidence of AI-based predictions is required in many real-world applications, such as biomedical assay development. My thesis focused on the development of a strategy called MIGHT, which I prove is guaranteed to quantify uncertainty and confidence given sufficient data. The key insight was that it is possible to integrate canonical cross-validation and parametric calibration procedures within a non-parametric ensemble method. Simulations demonstrate that while typical AI based-approaches cannot be trusted to obtain the truth, MIGHT can be. I applied MIGHT to answer an open question in liquid biopsies using cell-free DNA in individuals with or without cancer: which biomarkers, or combinations thereof, can we trust? Surprisingly, we find that combinations of variable sets often decrease rather than increase sensitivity over the optimal single variable set - because some variable sets add more noise than signal. Our work demonstrates the importance of quantifying uncertainty and confidence — with theoretical guarantees — for the interpretation of real-world data.
We next apply MIGHT to the analysis of cell-free DNA and circulating proteins in individuals that are of high-risk for being diagnosed with cancer. We performed shallow whole-genome sequencing (~1x) on the cell-free DNA of 1,110 plasma samples from 1,051 individuals with that were diagnosed with no disease, autoimmune disease, vascular disease, or cancer. We developed a comprehensive metric, called fragmentation signatures, that integrated the distributions of fragment positioning, fragment length, and fragment end-motifs. Using this metric, we found that individuals with venous thromboembolism, systemic lupus erythematosus, dermatomyositis, or scleroderma had cfDNA fragmentation signatures that closely mimicked those found in individuals with advanced cancers. Furthermore, these signatures were highly correlated with increases in inflammatory markers in the blood. Though these data put substantial limitations on the specificity of fragmentomics-based tests for cancer diagnostics, they also offer ways to improve the interpretability of such tests. Moreover, they should lead to a better understanding of the cells – most likely inflammatory cells - from which plasma cfDNA is derived. Here, I will describe findings that initiated serendipitously demonstrating these patterns are not specific to cancer patients and can arise in the absence of any neoplastic cells