22689 research outputs found
Sort by
ADVANCING UNDERSTANDING OF THE ETIOLOGY AND EARLY DETECTION OF SMOKING-ASSOCIATED AND OTHER CANCERS USING ‘OMICS MARKERS
No full textCigarette smoking is the leading cause of 13 types of cancer, with one in three cancer deaths attributed to smoking in the US. Current screening guidelines for smoking-associated cancers rely on self-reported smoking history (i.e. status, pack years, when to quit) or do not consider smoking history to identify high-risk individuals for screening. Nevertheless, self-reported smoking history may not be sufficiently accurate for quantifying varying personalized risk even among individuals with the same cumulative smoking history. Other sources of smoke exposure beyond active smoking, such as prenatal smoking and secondhand smoke, are also not reflected in self-reported smoking history. We believe a novel blood-based tool for early cancer detection can improve screening for smoking-related and other major cancers. The objectives of this thesis were to 1) refine the risk assessment of smoking-related cancers with methylation markers for smoke exposure in uterus (Aim 1) using data from CLUE I and II cohort, and proteomic-derived markers for smoke exposure in adulthood (Aim 2) using data from the Atherosclerosis Risk in Communities study (ARIC), 2) develop a new tool for early cancer detection based on the proteomic markers (Aim 3) in ARIC. We found that prenatal smoking predicted by DNA methylation markers was associated with increased lung cancer risk, where the association was the strongest in the 1930-38 birth cohort (Aim 1). The protein-based score reflecting personal smoking history was significantly associated with incidence of and mortality from smoking-associated cancer and lung cancer and can further provide additional relative risk information for smoking-associated cancers beyond self-reported smoking status and packyears and other known cancer risk factors (Aim 2). We identified most informative prediagnostic proteins which provided adequate sensitivity and specificity for the prediction of all cancer in the asymptomatic population. Though further optimization will be needed for population implementation for screening (Aim 3). Taken together, this thesis will inform risk stratification for screening of smoking-related cancers regarding prenatal smoke exposure and individualized adult smoking effects and development of early cancer detection tools, which can potentially guide future strategies for cancer prevention or risk mitigation
Essays on the Regulation of Modern Markets
No full textThis dissertation comprises three essays that examine the impact of government regulations on firm strategies and consumer behavior in modern markets.
The first essay studies exclusive dealing practices in China's online food delivery platform market and analyzes competition and welfare effects of banning exclusive dealing. I develop a model of the food delivery market to investigate restaurants' strategies in platform participation and pricing, as well as the subsequent consumer demand for meals. The model is estimated using a comprehensive dataset that covers a period when Eleme and Meituan, the dominant platforms in China's food delivery market, used exclusive dealing practices. The estimation results highlight substantial heterogeneity in consumer preferences for platforms and reveal high fixed costs for merchants joining multiple platforms. The counterfactual analysis indicates that banning exclusive dealing incentivizes merchants to join multiple platforms, though not necessarily leading to overall welfare gains. The findings highlight the challenges in formulating effective regulations of exclusive dealing.
The second essay investigates the rebalancing strategies of a bikeshare system operator under government equity-oriented incentives. My model studies how the operator chooses bike rebalancing frequencies for different service areas so as to maximize their profit, considering rebalancing costs and the impact of station overstock or understock on the variable profit and equity-oriented incentive payments from the government. Utilizing data from Divvy, Chicago's station-based bikeshare system, the study finds that the operating costs depend significantly on user trips and rebalancing efforts. The counterfactual analysis suggests that the city’s current incentive scheme motivates frequent rebalancing and reduces occurrences of station overstock and understock.
The third essay studies the introduction of e-scooter systems alongside bikeshare systems, focusing on Chicago's 2022 shared e-scooter program. Using a recent multi-period difference-in-differences method, I find that the overall bikeshare ridership remained unaffected by the introduction of e-scooter services. However, bikeshare usage among Divvy members significantly increased in areas where the same operator, Divvy, also provided e-scooter services. These findings suggest potential complementarities between Divvy's e-scooter and bikeshare services for members, which do not extend to the non-member population. The results could offer valuable insights for regulators overseeing this market
Interaction of Poly(ADP-ribose) with PARP13 Regulates Stress Granule Maturation and Dynamics
No full textPoly(ADP-ribose) (PAR) forms condensates through non-covalent interactions. The
influence of these interactions on the properties of cellular condensates is unclear,
primarily due to the dearth of identified molecular interactions between PAR and protein.
In this study, we demonstrate that PAR-mediated interactions through PARP13.2
modulate the dynamics and physical properties of stress granules. Expression of a
PARP13.2 mutant with reduced PAR-binding activity results in the assembly of smaller
but more numerous stress granules. This PAR-binding mutant diminishes both stress
granule dynamics and fusion. The fragmented stress granule phenotype of PARP13.2
mutant expressing cells is conserved across a variety of stresses that trigger stress
granule formation via diverse pathways. In addition, this phenotype is apparent in cells
expressing PARP13.2 variants with cancer-associated SNPs that disrupt PAR binding.
Overall, our study uncovers the important role of PAR-PARP13.2 interactions in
modulating stress granule dynamics and maturation, which provides new perspectives on
potential disease mechanisms
CROSS-MODAL PLASTICITY IN CORTICAL REGIONS FOLLOWING VISUAL DEPRIVATION
Get PDFThe brain has an amazing capacity to adapt to perturbations that affect the proper flow of information. One such perturbation, sensory deprivation, leads to the widespread plasticity of cortical circuits thought to underpin the compensatory changes that allow the brain to adjust. Visual deprivation elicits these circuit-level changes in the deprived primary visual cortex (V1) as well as the spared cortical regions like the primary auditory cortex (A1), the primary somatosensory cortex (S1), and the retrosplenial cortex (RSP). Here we address the role of corticocortical inputs in mediating visual deprivation-induced plasticity of V1 that may mediate cross-modal recruitment. We find that amongst the corticocortical inputs that bring multisensory information to V1, only inputs from the higher-order visual area, lateromedial area (LM), potentiate as a result of visual deprivation. This provides an avenue for the cross-modal recruitment of V1. Concurrently, using an activity-dependent reporter line (cFos-TRAP2;Ai14), we examined how these adaptive changes affected activity across various brain areas following vision loss. In the cortical regions assessed, we find that V1 adapts to vision loss with an initial reduction and subsequent rebound in activity following prolonged deprivation. In lateral visual areas, there is a reduction in activity with no accompanying rebound. In the active cell population following vision loss, there is a shift in the laminar distribution towards superficial depths in V1 and lower depths in lateral visual areas. In auditory areas, there was a trend toward a decrease in activity and A1 showed a relative increase in active cells at depths farther from the pia. Somatosensory cortices and the retrosplenial cortex show a significant reduction in the density of active cells but without redistribution across different depths. Altogether, our results highlight a potential avenue for the recruitment of V1 following visual deprivation and how cortical regions change their activity to adapt to this loss
Working Hard for the Money: Students' Entry and Exit Experiences with College Financial Aid
Get PDFWith the soaring costs of college, students (especially those from historically disadvantaged backgrounds) often enter higher education without adequate financial resources to support their experiences. To bridge this gap, students work through the Free Application for Federal Student Aid (FAFSA), Pell Grant requirements, and work-study jobs to reach the postsecondary promises of social mobility. But what is the right choice? Which approach to financing college, if any, pays off for students? My three-article dissertation quantitatively analyzes the extent to which students are securing (entry) enough financial resources to succeed (exit) in college. I examine 1) the role high school counselors play in getting students to file the FAFSA, 2) the problems and possibilities of the Pell program, and 3) how hours on the job during college affect academic achievement. The project is timely, as research and reform efforts are mostly focused on student loans and are based on data from before the Great Recession. By testing causal inference models on the High School Longitudinal Study (N = 25,206), my dissertation signals which types of work aspiring students should invest in to pay for college. In other words, what work pays off: filing financial aid forms, getting need-based grants, or setting up employment arrangements? While the answer to this question is vital for all, it is particularly important for ethnically and economically diverse students. The results of these analyses document the tradeoffs for students with respect to each approach to financing their education. In addition to helping students navigate the financial aid process, this dissertation will be useful in helping university, state, and federal policymakers evaluate their strategies for improving college affordability
SEVENTEENTH-CENTURY HISTORICAL PERSPECTIVES ON THE THEORY OF DIVINATION
Get PDFThis dissertation examines the early modern historiographic idea that divination originated in the pre-classical societies of the Middle East. It traces a direct connection from debates about ancient authority to scholars’ disavowal of astrological and magical types of foresight as subjects of scientific interest. It does so in the context of debates that took place in the 17th-century French Republic of Letters, an informal, influential network of humanists and natural philosophers which enabled the long-distance exchange of materials and ideas through correspondence. I address how scholarship in this community drew on astrology, biblical studies, Christian Kabbalah, and travel writing to probe the narrative that the origins of science and philosophy were described in sacred history and disseminated in Hebrew, Chaldean, and Egyptian sources. Research produced in the circles of polymaths Nicolas-Claude Fabri de Peiresc, Pierre Gassendi, Marin Mersenne, and Ismaël Boulliau frequently addressed the reality of divination from its questionable historical origins, rather than its internal merits. These studies became broader and more sophisticated as France’s commercial and diplomatic position in the Near East became entrenched in 1620-1650. I show how these correspondents debated the historicity of ancient magic in the context of antiquarian manuscript and material studies, Franco-Ottoman diplomacy and travel, and Hebraist scholarship. As the knowledge European humanists acquired through these channels diversified, a methodological shift occurred and writers in this environment became more skeptical about the epistemic possibilities of history. Contemporary stereotypes of Middle Easterners’ superstitiousness threatened to undermine Christian humanists’ religious commitments, reducing confidence that these sources could produce novel insight into the ancient philosophical and theological tradition. Scholars increasingly tried to understand Near Eastern sources in terms of their immediate socio-cultural context, rather than their absolute veracity. When Renaissance scholars no longer felt they needed to filter intellectual history through a narrative of diffusion from philosophical sages, the Republic of Letters could no longer support a dialogue between pro- and anti-divination positions. In the end, the decline of divination was caused by an ideological shift in prevailing historiographic conventions, not by scientific debate
TO HEAL IS TO LIVE A YOUNG WOMAN’S PERSPECTIVE ON HEALING WITHIN THE MIND, BODY, AND NATURE
No full textHealing can be a slow, painful process. Sometimes it can be as arduous as the trauma that impacted the mind, body, or soul in the first place. Writers have the honor, and challenge, of telling stories of triumph and those of triumph in the making. In my twenty-seven years of life, I have experienced death first-hand, observed people face insurmountable systemic challenges, and seen land erode before my eyes. At the same time, I’ve seen men regain a long-lost act of pride and women reconnect with the vastness of the universe
Parametrically upscaled coupled constitutive damage model for piezoelectric composites
No full textPiezoelectric composites, known for their strong electromechanical coupling capabilities, are used in various engineering applications requiring sensing and actuation functions. These composites offer improved performance compared to monolithic piezoelectric ceramics. However, owing to heterogeneous microstructures, these composites are susceptible to different failure modes at the micro-scale, ultimately impacting their electromechanical performance at the macro-scale. Consequently, damage in these composites is a multi-scale phenomenon and their macroscopic response is strongly influenced by the underlying microstructure and the physics governing its deformation. Phenomenological constitutive models used for macroscopic analysis often ignore the effect of the underlying material microstructure and its physics, resulting in inaccuracies. Although hierarchical multi-scale models are available, they are computationally expensive for nonlinear electromechanical damage problems. To overcome these shortcomings, a computationally efficient microstructure-sensitive parametrically upscaled coupled constitutive damage model (PUCCDM) for piezoelectric composites is developed in this dissertation.
To realize the effects of the microstructure at macro-scale, it is imperative to understand different failure mechanisms in heterogeneous piezoelectric composites at the micro-scale. Towards this goal, the first half of the dissertation develops a finite deformation cohesive zone enhanced phase field model to simulate electromechanical fracture in these composite microstructures. Fracture in these composites is driven by complex mechanical and electrical mechanisms governed by a set of strongly coupled nonlinear governing equations. The model incorporates cohesive traction-separation laws at material interfaces via an auxiliary phase field order parameter. A Gibbs free energy density function, considering the anisotropic elastic stiffness of piezoelectric materials is proposed. Numerical simulations showcasing different failure mechanisms are carried out to demonstrate the model's efficacy.
The later part of the dissertation concentrates on the development of the PUCCDM, a higher-scale multi-physics damage model explicitly accounting for the influence of material microstructure. The microstructure comprises of nonuniformly distributed unidirectional piezoelectric fibers embedded in an epoxy matrix. Key microstructural features influencing macroscopic responses are identified and expressed as representative aggregated microstructural parameters (RAMPs), and statistically equivalent representative volume elements (SERVEs) are constructed. A database of SERVEs is created and simulations are performed under different loading conditions to calibrate the PUCCDM constitutive parameters as functions of RAMPs using machine learning tools. The developed PUCCDM exhibits computational efficiency, making it an indispensable tool for conducting multi-physics and multi-scale analysis of damage in multifunctional composites thereby enabling the material-by-design process
Disease Shocks and Power in the International System
No full textThough large-scale outbreaks of infectious disease recur throughout history, their effects on the
international system have been understudied. This dissertation is an attempt to understand how sudden,
large-scale outbreaks of acute infectious disease may affect power relations between states in the
international system. It proposes a model through which “disease shocks” affect the material and social-organizational bases of state power. Changes in state power affect relations between states in the
international system. The dissertation first considers the nature of disease shocks, then the nature of
disease effects on state power, and subsequently the nature of disease effects on the international
system. It then focuses on historical cases of asymmetric outbreaks that demonstrate the potential of
disease to affect power relations between states. Finally, it considers the potential use of disease for
strategic purposes, and what states might do to face such a challenge. This dissertation inserts sudden,
large-scale outbreaks of acute infectious disease into existing frameworks for understanding power
relations in the international system, builds a skeletal framework that supports further theoretical and
empirical work, and brings forward the strategic problem of disease shocks, opening the way to additional
research as well as policy development in the future
Integrating Physics-Based and Data-Driven Models for Task Autonomy in Robotic Eye Surgery
No full textAutonomous surgery holds immense potential for aiding surgeons in precision-driven tasks, enhancing safety, and streamlining complex surgical procedures. These advancements can lead to improved surgical outcomes, reduced surgeon errors, and broader access to care. However, most current robotic platforms remain teleoperated, serving a passive role during surgeries. This leads to the question: how can we move beyond building assistive systems towards intelligent autonomous systems? How can these systems be designed to automate parts of the surgery, while ensuring efficacy, safety, and minimally impacting the surgical workflow?
This dissertation seeks to answer these question in the application domain of retinal surgery. The overarching objective is to develop intelligent systems that shift from mere assistive roles to more autonomous ones. An incremental realization of such a goal, for instance, is a system that performs a subtask of a long procedure autonomously, while a surgeon provides close oversight. Towards this end, we leverage various multi-modal data, including robot kinematics, color and depth images, and force-sensing, to synthesize autonomous surgical behaviors which has largely been unexplored in prior work.
A crucial insight in developing such systems is the fusion of physics-based and data-driven models in a manner that maximizes efficacy and safety. Data-driven models offer the advantage of addressing complex robotic problems in a relatively convenient, model-free manner. However they are prone to error in novel scenarios and may not be ideal for meeting hard constraints concerned with the safety of the patient. By integrating these data-driven models with physics or logic-based approaches, which excel in adhering to strict safety criteria, we achieve a complementary safety element. In essence, we combine the best of both worlds of physics-based and data-driven models. This insight underscores the need to discern the strengths and limitations of available tools in order to synthesize a hybrid solution tailored appropriately to the specific challenge at hand. We illustrate that the needs of the application often dictate how these approaches may be integrated.
To demonstrate the efficacy of such systems, we first begin with a deep dive into the fundamental task of autonomous surgical tool navigation, which is arguably the simplest yet among the most difficult task to accomplish accurately in retinal surgery setting. Then, we explore more downstream applications including autonomous vein cannulation and autonomous subretinal injection to demonstrate a significant automation of these procedures. Thorough experiments with cadaveric pig eyes are demonstrated, and strong results demonstrate potential extension of these works for live animal studies. The proposed methods and solutions, however, are not without limitations, and directions for future work are suggested