71446 research outputs found
Sort by
Beyond the Stars: Optimizing Residual Images and Source Counts in Deep-Field Spitzer IRAC Mosaics
No full textThis thesis presents refined approaches for the development of source count catalogs and residual images for deep-field Spitzer Infrared Array Camera (IRAC) mosaics. Two well-known scientific fields, the Euclid Deep Field North (EDF-N) and the Euclid Deep Field Fornax (EDF-F), are analyzed at 3.6 microns and 4.5 microns, resulting in a final set of four residual images and source counts. The residual images are developed through an iterative point spread function (PSF) fit-and-subtract routine that is validated on the basis of normalized completeness and photometric accuracy. The source counts are cataloged as an additional component of the iterative subtraction method and evaluated for precision and completeness through the use of SourceXtractor++ and published historical IRAC source counts. Three of the four sets of final residual images and source count catalogs are well-suited for scientific use, specifically as they relate to the study of the cosmic infrared background and its broader cosmological implications.Applied Mathematic
Olmsted in Milwaukee: A Comparative Ethnography of Lake and Riverside Parks
No full textLake Park and Riverside Parks have a long history in Milwaukee from their creation to the present. In 1891, Frederick Law Olmsted designed these parks as part of a larger park system that was connected via the Newberry Boulevard greenway. Over a century later, Lake Park has retained most of its original elements and remains true to Olmsted’s original intention for the park, however, Riverside Park has drastically changed over the decades and has largely fallen into disrepair. This project aimed to investigate the reasons why Riverside Park has undergone significant changes while the structure of Lake Park has remained largely the same.
This project explores the literature pertaining to ecological anthropology and how it relates to each of these parks over time. Archival research was performed to explore Olmsted’s history as a landscape architect and examine his philosophy as it pertains to values and principles of park design. Extensive archival research was conducted to obtain historical information for each park and historical photographs were examined to make comparisons to the current conditions of specific park elements, and the Toolkit for the Ethnographic Study of Space (TESS) was used to identify the ways in which Lake and Riverside Parks were used. Ken Leinbach was interviewed and provided information on the history of Riverside Park and the Urban Ecology Center.
Findings indicated that Lake Park was primarily used for leisure activities, sports and recreation, as well as being a site for larger city events, whereas Riverside Park was primarily used as a site for nature-based activities, whether they be educational or
recreational. This project resulted in the ironic and revelatory finding that Riverside Park has surprisingly become a more accurate reflection of his philosophy and more closely represents vision. While the structure of Lake Park has remained truer to Olmsted’s original landscape design, Riverside Park has become a more accurate representation of that which Olmsted values in his philosophy of park design.Extension Studie
Neural Stem Cell (NSC) Secretome and Wnt/β-catenin Signaling Pathway in Alzheimer’s Disease (AD): A Systematic Literature Review and Bioinformatics Analysis
No full textAlzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by neuronal loss, synaptic dysfunction, and cognitive impairment, with no current cure (DeTure & Dickson, 2019; Kumar et al., 2023). The Wnt/β-catenin signaling pathway plays a crucial role in neurogenesis and synaptic plasticity; however, its dysregulation in AD, due to pathological factors such as amyloid-β (Aβ) plaques and neurofibrillary tangles (NFTs), contributes to disease progression (Jia et al., 2019; Priya et al., 2024). Neural stem cell (NSC) secretome, a collection of bioactive molecules secreted by NSCs, has shown potential in activating the Wnt/β-catenin signaling pathway and promoting neurogenesis in AD (Bahlakeh et al., 2022; Hijroudi et al., 2022). However, the specific components within the NSC secretome that contribute to Wnt/β- catenin activation and their underlying mechanisms remain unknown. This study aimed to identify the components of the NSC secretome, specifically growth factors and neurotrophic factors, that are responsible for activating the Wnt/β-catenin signaling pathway in AD, thereby enhancing neurogenesis, and to investigate their potential mechanisms of action through a comprehensive literature review and bioinformatics analysis. A systematic literature review identified eight growth factors – epidermal growth factor (EGF), fibroblast growth factor 2 (FGF2), fibroblast growth factor 8 (FGF8), hepatocyte growth factor (HGF), insulin-like growth factor-1 (IGF-1), platelet- derived growth factor (PDGF), stem cell factor (SCF), vascular endothelial growth factor (VEGF) – and three neurotrophic factors – brain-derived neurotrophic factor (BDNF), glial cell line-derived neurotrophic factor (GDNF), nerve growth factor (NGF) – within the NSC secretome that may contribute to Wnt/β-catenin activation in AD. Bioinformatics analysis categorized these proteins using UniProt database, assessed their molecular and biological functions through Gene Ontology (GO) analysis using PANTHER database, and identified their pathway involvement through KEGG database. STRING-based protein-protein analysis (PPI) network analysis provided insight into the intricate crosstalk between PI3K/Akt, MAPK/ERK, and Wnt/β-catenin signaling pathways. Then a targeted literature review was conducted to validate bioinformatics predictions, identifying two direct and one indirect crosstalk mechanism of Wnt/β-catenin signaling pathway activation. The first direct mechanism involves growth factors and neurotrophic factors activating the PI3K/Akt signaling pathway, leading to Akt-mediated phosphorylation and inhibition of glycogen synthase kinase-3β (GSK-3β), which stabilizes β-catenin and promotes its nuclear translocation to activate Wnt target genes. The second direct mechanism involves EGF binding to epidermal growth factor receptor (EGFR) triggering MAPK/ERK activation, where extracellular signal-regulated kinase (ERK) phosphorylates casein kinase 2 (CK2), which subsequently phosphorylates α- catenin. This phosphorylation event disrupts the β-catenin-E-cadherin complex, allowing β-catenin to dissociate and translocate into the nucleus to drive Wnt target gene transcription. The indirect mechanism involves MAPK/ERK signaling activation by growth factors and neurotrophic factors, where rat sarcoma (Ras) activation converts phosphatidylinositol 4,5-biphosphate (PIP2) to phosphatidylinositol (3,4,5)-trisphosphate (PIP3), reinforcing PI3K/Akt pathway activation and enhancing β-catenin stabilization. A final pathway interaction map was constructed to visualize the complex interplay between these signaling pathways and their potential therapeutic implications in AD. Collectively, these findings highlight the potential of NSC secretome components in restoring Wnt/β-catenin signaling in AD through intracellular crosstalk with PI3K/Akt and MAPK/ERK signaling pathway. This study provides a molecular framework for future exploration of NSC secretome-based therapies in AD.Extension Studie
Investigating how sequence variation in T cell receptors and human leukocyte antigens shapes T cell development
No full textT cells are critical agents of the adaptive immune system that orchestrate killing of infected and cancerous cells. Some T cells erroneously target healthy cells, leading to autoimmune disease. Whether a T cell becomes activated depends on the immunological synapse it forms with an antigen presenting cell (APC). At the core of the immunological synapse are two key proteins: (1) the T cell receptor (TCR) and (2) the major histocompatibility complex (MHC) presenting a peptide from a candidate pathogen.
This dissertation examines genetic variation on both sides of this immunological synapse. Genetic variation in the MHC confers extreme risk for autoimmune disease and largely consists of germline polymorphisms, which are consistent across an individual’s cells. In contrast, genetic variation on the T cell side of the immunological synapse is largely somatic. Each developing T cell stochastically rearranges and edits the genes encoding its TCR, creating a repertoire of TCR sequences within each individual.
Analysis of TCR sequence data traditionally proceeds by considering each TCR as a “molecular barcode” and identifying T cells that match exactly on this molecular barcode. However, the vast majority TCR sequences are observed only once, in one individual.
A major theme throughout this dissertation is an alternate approach to TCR sequence analysis: (a) decomposing each TCR amino acid sequence to a collection of physicochemical features such as hydrophobicity and electrostatic charge, and (b) testing how these sequence features relate to T cell development outcomes, such as thymic selection and transcriptional fate. This approach enables two general discoveries: (1) TCR sequence features regulate T cell transcriptional fates, and (2) MHC genetic variants which confer risk for autoimmune disease influence which TCR sequences pass thymic selection. Specifically, we find that hydrophobicity throughout the CDR3 region of the TCR sequence promotes regulatory T cell fate. Most surprisingly, we observe a constellation of TCR sequence features that are consistently enriched in memory T cells compared to naïve T cells. We develop a TCR scoring function “TCR-mem,” which quantifies the extent of these features in each TCR, and show through TCR transduction experiments that increased TCR-mem increases T cell activation even among T cells that recognize the same peptide-MHC complex. Each of these projects provide new scoring functions that allow researchers to score and rank TCR sequences for functional follow up, regardless of whether the TCR sequences have been observed previously.
Altogether, this dissertation provides new approaches to study sequence variation in MHC and TCR molecules, and sheds light on how this variation may alter critical T cell functionality.Biomedical Informatic
The Big One: Understanding Risk in an Uncertain World.
No full textIn today’s unstable nuclear multipolar world, issuing threats —both physical and verbal—has become central feature of strategic interactions between nuclear-armed states. "In a nuclear world filled with noise, policymakers often struggle to discern which threats are real and require urgent attention, and which are merely posturing or signaling."
This thesis develops and explores a risk-based framework that might help policy makers gain much needed clarity. The framework is applied in this thesis specifically to decipher threats that the Russian Federation has emanated from 2014 to 2024 to determine which are the threats that require focus.
This framework draws from risk assessment models and resilience-based regulation in the business world to identify threats, assess situations, and determine appropriate responses. It evaluates whether the threat places the current status quo at risk, or if the status quo has been redefined, potentially increasing levels of risk. Finally, it identifies the triggers necessary for re-evaluation—an essential process for understanding, quantifying, and managing business risks.Extension Studie
Proof of Life: The Biopolitics of Visual Media
No full textThis study outlines a genealogy of media that claim to prove something about human life. It draws on archival and historiographical research to analyze interactions between photography, the history of statistics, activist art and video, public health crises, and digital media. Across a series of case studies spanning the twentieth and twenty-first centuries, the dissertation argues that visual media emerges as a biopolitical tool for governing populations in tandem with new forms of biometric surveillance that work to structure space and movement. The chapters examine the relationship between photography, social reform, and statistics; the role of aesthetics in AIDS activist media, art, and public health campaigns; the history of biosurveillance in relation to the cultural and political imaginaries that emerged after the completion of the Human Genome Project; and the role of contemporary artists in understanding how the body is reconceptualized—or reformatted—as an object for the extraction of data within regimes of “surveillance capitalism.” The concluding chapter looks to early experiments with responsive environments in artificial intelligence research to contextualize productive encounters between statistical science, cybernetics, architectural design, and organizational management in the mid-twentieth century that implemented multimedia surveillance as a mechanism of control—encounters that continue to influence the ideological aims and technical operations of contemporary therapeutic and assistive digital tools. By focusing on the diagnostic aspect of biometric capture and visual surveillance in relation to risk management, the dissertation frames media as an unexpected technique and tool of biopolitical governance. Across chapters, Proof of Life: The Biopolitics of Visual Media introduces the concept of "biosurveillance media"—media forms designed to surveil, track, and quantify the body. The dissertation contributes to a critical body of scholarship in film and media studies and the history of technology that unravels the fraught relationship between media technologies, political agency, and the governance of human life.Film and Visual Studie
Essays on Consumer Preferences in Health
No full textThis dissertation quantifies consumer preferences in the domain of health, examines the mechanisms underlying these preferences, and develops tools for policymakers and practitioners to incorporate these insights into health policy decision-making. Chapter 1 provides evidence that people's ability to process complicated numerical information influences their willingness to pay for small changes in the risk of dying, potentially biasing economic evaluations of public policies that reduce health risks. To address this challenge, this chapter develops a statistical approach that practitioners can use to correct for bias, improving estimates of the value of public policies. Next, Chapter 2 (co-authored with Angelique Acquatella and Amitabh Chandra) examines whether people's preferences for resource redistribution depend not only on beneficiaries' financial status but also on their health status. It finds that many people base their distributional preferences on both health status and income, and it develops conceptual methods to aggregate these preferences, which can guide the future evaluation of social policies that affect individuals with health conditions. Finally, Chapter 3 documents diverse public views on whether health insurance plans should charge higher prices to specific consumer groups that may have higher resource utilization, highlighting the important role of people's beliefs about personal responsibility for health. Documenting this heterogeneity helps policymakers to create policies that people will perceive as fairer and will be more likely to support. Together, these chapters demonstrate the importance of understanding public perceptions of value and equity in designing and evaluating health policies.Health Polic
Training Energy-Based Models to Learn Gaussian Mixture Distributions via Langevin Dynamics
No full textEnergy-Based Models (EBMs) offer a flexible framework for modeling probability distributions through an energy function, avoiding the need for explicit density functions. Using toy distributions such as Gaussian Mixture Models (GMMs). This thesis examines variations of Langevin Dynamics as a method for sampling and training EBMs, providing concrete implementations. We present the theoretical foundations and practical implementations of key EBM training algorithms, including Unadjusted Langevin Algorithm (ULA), Metropolis-Adjusted Langevin Algorithm (MALA), and adapting the Jarzynski Equality for sampling and training. Results demonstrate the benefits and limitations of these methods, highlighting how while the Jarzynski Equality offers a time-dependent perspective that accelerates the discovery of distribution modes, it introduces practical challenges, such as difficulty of implementation and instability during training. Future work involves refining the Jarzynski-based training procedure, exploring alternative loss functions like Fischer Divergence, and extending the compositional capabilities of EBMs for tasks in robotics and reinforcement learning.Computer Scienc
MITOCHONDRIAL TRANSPLANTATION AS A NOVEL THERAPEUTIC STRATEGY FOR RETINAL ISCHEMIA-REPERFUSION INJURY
No full textAbstract
Title: Mitochondrial Transplantation as a Novel Therapeutic Strategy for Retinal Ischemia-Reperfusion Injury
Authors: Fernando Rubio-Mijangos1,2, Aybuke Çelik1, Maria Loscertales2, Alexander Bigger-Allen3, Sitaram Emani1, Pedro del Nido1, Demetrios Vavvas2*, James D. McCully1*
1Department of Cardiac Surgery, Boston Children’s Hospital, Department of Surgery, Harvard Medical School, Boston, MA, 02215, USA
2Department of Ophthalmology, Retina Service, Massachusetts Eye and Ear, Harvard Medical School, Boston, MA 02114, USA
3Urological Diseases Research Center, Boston Children’s Hospital, Harvard Medical School, Boston, MA 02114, USA
*Co-last authors
Keywords: Mitochondrial transplantation; Retinal ischemia-reperfusion injury; Oxidative stress; Mitochondrial dysfunction; Cellular bioenergetics; ARPE-19; Inflammation; Apoptosis;
Background
Mitochondria are central to cellular homeostasis, providing ATP through oxidative phosphorylation while also regulating oxidative stress, apoptosis, and immune responses. Retinal ischemia-reperfusion injury (RIRI) disrupts mitochondrial function, triggering oxidative stress, inflammation, and neurovascular degeneration. Given the critical role of mitochondria in retinal cell survival, mitochondrial transplantation (MT) has emerged as a promising therapeutic strategy to restore mitochondrial integrity and bioenergetics in ischemic tissues. This study evaluates the potential of MT in rescuing oxidative stress-induced mitochondrial dysfunction in an in vitro model of RIRI.
Methods
ARPE-19 cells were differentiated into a mature retinal pigment epithelium-like phenotype and subjected to oxidative stress using hydrogen peroxide (H₂O₂). Mitochondria were isolated from ARPE-19 cells and characterized for viability and function. Transplantation was performed at optimized doses, and mitochondrial uptake, ATP production, oxidative stress resilience, and cell survival were assessed. Bulk RNA sequencing was conducted to analyze transcriptional responses following MT, identifying differentially expressed genes and enriched biological pathways involved in cellular stress adaptation.
Results
Transplanted mitochondria were efficiently internalized by ARPE-19 cells and integrated into the host mitochondrial network, restoring ATP levels in a dose-dependent manner. MT significantly improved cell survival and reduced oxidative stress-induced apoptosis at 0.25-0.5 mM H₂O₂ concentrations. RNA sequencing revealed that MT downregulated inflammatory and apoptotic pathways, including IL-17, VEGF, and CASP9 signaling, while upregulating stress-adaptive pathways such as TNF, NF-κB, SGK1, and efferocytosis. Notably, MT reduced IL6 and IL6R expression, potentially mitigating inflammation-driven mitochondrial dysfunction.
Conclusion
MT demonstrates significant potential in restoring mitochondrial function and modulating stress-responsive transcriptional programs in RIRI models. By mitigating oxidative stress, preserving ATP levels, and reprogramming inflammatory responses, MT presents a promising therapeutic approach for retinal ischemic diseases.
Future Directions
Further investigations are required to validate these findings in vivo, refine transplantation protocols, and assess the long-term efficacy and safety of MT. Proteomic and metabolomic analyses could provide deeper insights into mitochondrial integration and functional recovery, advancing MT as a viable intervention for retinal ischemia.Graduate Educatio
Cooperative Multi-Agent Graph Bandits
No full textThis thesis studies multi-agent, sequential decision making under restrictions on consecutive actions by extending the multi-armed bandit (MAB) paradigm, a fundamental online learning framework with widespread applications in healthcare, recommendation systems, dynamic pricing, and generative artificial intelligence. Specifically, we present the multi-agent graph bandit problem in which N cooperative agents navigate a connected graph G with K nodes. At each time step, agents play an action associated with their current location and observe a random reward. The edges of G thus represent restrictions on which actions can be played sequentially. Unlike in the existing multi-agent MAB literature, we define a coupled reward function, with the total reward of the system formulated as a weighted sum of the rewards sampled by individual agents.
To address this novel learning scenario, we present the Multi-G-UCB algorithm, which episodically estimates, transitions to, and samples from the multiset of nodes that give optimal system-wide reward. We bound our algorithm's regret (a traditional performance measure in online learning settings) by O(γ√(NKT log T) + γDNK log T), where T is the time horizon, D the diameter of G, and γ a boundedness parameter associated with the weight functions. Our agent-average regret bound is tighter than that of any single-agent algorithm and, when T is large, our collective regret matches state-of-the-art bounds from related but simpler bandit formulations.
Motivated by challenges introduced in real-world deployments, we generalize our framework to model agent and communication failures and develop a robust implementation of our algorithm, Robust-Multi-G-UCB, that achieves the same asymptotic performance in the robust setting. We empirically demonstrate that our algorithm surpasses several important benchmarks and performs well in a diverse array of challenging problem instances before finally recording a hardware demonstration deploying Robust-Multi-G-UCB in real-time on a swarm of light-sensing TurtleBots.Computer Scienc