80264 research outputs found
Sort by
Bayesian Modeling of Multivariate Mixed Data
No full textBayesian methods provide a powerful framework for analyzing complex multivariate
data, particularly when dealing with dependencies among variables and with mixed
type of data, which may include continuous, categorical, and count variables. In many
real-world applications, data exhibit dependency structures that standard modeling
approaches struggle to capture effectively. Developing novel methodologies that can
capture these relationships while accommodating mixed-type data is essential for
improving inference and prediction. In this thesis, we introduce three novel Bayesian
approaches for multivariate mixed data, addressing challenges in modeling count data,
mixed graphical models, and varying coefficient logistic regression modeling.
First, we present a Bayesian bivariate Conway-Maxwell-Poisson (CMP) regression
model for analyzing correlated count data in sports. This model provides a flexible
framework that accommodates over-, under-, and equi-dispersed count data, yielding
improved fitting and more accurate estimations compared to standard Poisson and
Negative Binomial models. The proposed approach can be extended to multivariate
settings and is applicable to a wider range of domains beyond sports analytics.
We demonstrate the effectiveness of our model through simulations and real-world
applications, focusing on two major sports: soccer and baseball.
Second, we propose a Bayesian approach for inference in mixed graphical models,
which enables the study of conditional independencies among mixed data, where
observed variables can take multiple forms, including continuous, categorical, count,
and zero-inflated count data. Our method employs a flexible conditionally specified
modeling framework that captures the underlying structure of each variable, allowing
for the imputation of missing data based on potential interactions among variables of
different type. We prove that the proposed model matches or outperforms existing
methods, particularly in scenarios involving missing data. We apply our approach
to real data collected from adolescents diagnosed with an eating disorder, revealing
intrinsic relationships among the variables during the treatment.
Finally, we extend our Bayesian mixed graphical modeling framework by incorporating
a varying coefficient structure. In this approach, we model the conditional
independencies among predictors while allowing a set of covariates to modulate their
effects on a binary response variable. This framework provides a more flexible and
interpretable representation of complex interactions. We show that our model outperforms
competing methods in predictive accuracy and correctly estimates varying
coefficients. We further apply it to real data from adolescents with eating disorders,
uncovering key relationships between predictors and covariates and their influence on
predicting bulimia
Development of Extracellular Matrix-Based Colloidal Inks for Cartilage Tissue Engineering
No full textAdditive manufacturing enables spatial control over bioactive molecules and cells to mimic native tissue architecture, but a lack of bioinks that balance biological relevance and printability limits its potential. Decellularized extracellular matrix retains native biochemical cues but suffers from poor mechanical stability, restricting its use in 3D printing. This work presents the development of composite colloidal inks using methacryloylated decellularized cartilage extracellular matrix nanoparticles blending with gelatin nanoparticles to improve both printability and biofunctionality. The resulting inks are shear-thinning, self healing, and UVcrosslinkable, enabling the fabrication of tunable 3D-printing scaffolds. These scaffolds supported human bone marrow mesenchymal stem cell chondrogenesis, evidenced by enhanced collagen deposition, upregulation of chondrogenic gene expression, and suppression of osteogenic markers expression without exogenous differentiation factors. This study also explored the use of machine learning approaches to predict the print quality of 3D printed poly(propylene fumarate) and to identify relationships between printing parameters and the print quality. Print speed and material composition had the greatest effect on scaffold quality. Additionally, this work examined printing consistency with colloidal inks. Unlike poly(propylene fumarate), the colloidal inks required real-time parameter adjustments to maintain print fidelity, likely due to pressure-induced phase separation. Overall, this research introduces a novel, biologically active, and customizable colloidal ink platform for cartilage tissue engineering and broadens understanding of print behavior in colloidal systems
Beyond Transit-Oriented Development: Rethinking Urban Expansion in Suburbia
No full textAmerican cities are stuck in a stagnant suburban development pattern that continually replicates itself outward. Many of the social, economic, and environmental crises society faces today are perpetuated through single-use development that prioritizes private, individualized life over collective urban identity. This focus on individualized life also leads to stagnation in the built form. Transit-Oriented Developments (TODs) are an exception to this development pattern, which allow for an increase in density and address some environmental challenges of suburbia, but do little to change overall development patterns.
This project both criticizes and builds upon TODs as a method of urban growth and seeks to produce a catalyst for change in the suburbs of American cities. It identifies a series of patterns within American suburbs and proposes strategies for transforming each identified suburban typology into a place that can grow and change over time. This project aimed to visualize a way for American suburbs to exit their current development patterns and develop an urban form capable of producing urban life and multimodal transportation. It does so through the identification of a series of typological transects within the suburbs, and proposes how these transects could each transform over time while forming an integrated city together. The project also proposes strategies for connecting these spaces and shows an adapted version of the plans it proposes in the context of Reston, a suburb of Washington in Northern Virginia
Lost Landscapes
No full textThe Sutro Baths, once a monumental bathing facility in San Francisco, represented an early vision of leisure and commodified nature, transforming the natural coastline into an artificial space for entertainment, consumption, and spectacle. Established at the turn of the 20th century, the Baths blended modern architecture with the burgeoning leisure economy, featuring diverse programs such as swimming pools, promenades, and exhibitions, all housed under vast glass vaults. Despite its destruction in 1966, the site has since been reclaimed by nature, transitioning from an engineered landscape to a wild scrubland. This project explores the intersection of urbanization and natural ecosystems in San Francisco, offering a critique of how colonization and exoticism shaped the city’s environment, notably through the introduction of non-native species like eucalyptus. In response to these tensions, the proposed architectural intervention aims to recover and reinterpret the lost landscapes of the Sutro Baths site. By embracing the organic, dynamic topography of the area, the design weaves together circular pavilions, gardens, and tidal pools that harmonize with the natural landscape rather than imposing upon it. This project reimagines the bathhouse as a site of ecological awareness and cultural reflection, integrating educational spaces, ecological restoration, and a sustainable leisure program to reconnect the urban population with the region’s complex natural history
Theory of Molecular Eigenmodes of NMR Relaxation for Enhanced MRI Contrast
No full textIn medical Magnetic Resonance Imaging (MRI), the Nuclear Magnetic Resonance (NMR) relaxation of 1H in water is used to probe the state of tissues. The contrast in these images is enhanced by Gadolinium-based contrast agents (GBCAs), which act by reducing the water 1H proton relaxation time. However, despite the long history of NMR, the molecular-scale processes involved in NMR and MRI remain poorly understood. The modeling and interpretation of relaxation physics still rely on severe assumptions, limiting how we interpret and utilize this technology. Moreover, existing commercial GBCAs have been shown to release toxic Gd(III) ions in the body, causing severe renal impairments and bioaccumulation of heavy ions in various tissues. All this underscores the necessity for (i) improved NMR relaxation models that provide clearer insights into the molecular-level processes governing relaxivity, and (ii) methods to assess and enhance the stability of MRI chelated ions to prevent ion release and increase patient safety. This will ultimately enable the development of newer, safer, and more efficient MRI contrast agents.
In this work, we present how the physics of NMR relaxation autocorrelation can be decomposed into contributions from “molecular eigenmodes,” which contain critical information about the structure and dynamics of the system. We show that these molecular eigenmodes arise from the solution for the diffusion propagator of the system, whose time evolution is described by an appropriate Fokker-Planck equation. Based on this principle, we develop a statistical mechanical framework that predicts the multi-exponential decay of the NMR relaxation autocorrelation function, eliminating some of the constraints imposed by mainstream NMR relaxation models. In Molecular Dynamics simulations, these eigenmodes can be accessed through a multi-exponential Padé-Laplace inversion of the NMR relaxation autocorrelation function.
We have also investigated the NMR relaxivity of MRI contrast agents using quantum and molecular simulations at physiological temperatures. Specifically, we performed first-of-its-kind molecular dynamics simulations of Gd(III) and Gadavist® (a commonly used MRI contrast agent) using highly refined polarizable models derived from ab initio calculations. Simulation results for both Gd(III)-aqua and Gd(III)-DO3A-butrol complexes at human body temperature were corroborated and validated by NMR relaxation dispersion measurements. At frequencies relevant to MRI, our simulations of NMR relaxivity show good agreement with experiments, without relying on free parameters to interpret the simulations. This stands in stark contrast to the extended Solomon-Bloembergen-Morgan (SBM) model, the primary framework for interpreting relaxation in MRI.
Finally, we present how to estimate the chemical stability of chelating agents used in MRI. Using the molecular Quasi-Chemical Theory (m-QCT), we assess short-range energetic contributions to paramagnetic complexes through ab initio calculations, while long-range solvation contributions are estimated using continuum solvent models. This technique allows for the calculation of the equilibrium constant between the chelated and free ion states. The results show good agreement with excess hydration free energies for Gd(III) and Mn(II)-aqua complexes —another paramagnetic species of interest in MRI. Our method can not only identify more stable chelates for MRI contrast agents but also has potential applications in emerging radiopharmaceutical therapeutics
Enhancing the Tunability of Multidomain Peptides: Composite Systems and Synthetic Glycosylation
No full textMolecular self-assembly, the process of molecules organizing themselves into hierarchical structures through non-covalent forces, is responsible for many of life’s most important processes. Inspired by nature, scientists have created synthetic, self-assembling systems to tackle a wide range of biomedical problems. One area that these systems have shown promise is in the field of drug delivery. The ability to control where and when a drug is delivered, i.e. spatiotemporal control, is an issue for many drugs but most notably for chemotherapeutics in cancer treatment. As chemotherapeutics do not discriminate between cancerous and healthy cells, many horrible side effects occur due to off-target toxicity. Self-assembled, nano-sized structures such as nanofibers made from peptides and particles made from lipids have been used to deliver drugs more efficiently, but drawbacks such as subpar release kinetics or short half-lives leave much to be desired.
The first part of this work explores the use of multi-component, or composite, systems made up of multiple self-assembling systems to gain finer control over drug release kinetics. We use MultiDomain Peptides (MDPs), a class of self-assembling peptide hydrogels, and liposomes to create drug delivery systems with tunable release kinetics based on peptide-lipid interactions. We show that electrostatic interactions can be used to tailor the release kinetics, and when tuned correctly, can result in delayed release kinetics longer than either system alone. We go on to use these composite systems in an aggressive oral cancer model and show that a single injection of our immunotherapeutic-loaded composite system is as effective as six separate injections of free immunotherapeutic.
The second part of this work looks to expand the synthetic toolbox of MDPs through glycosylation. We show, for the first time, the use of MDPs in organic solution phase reactions, resulting in a library of eleven novel glycosylated MDPs. We investigate the effects of position, frequency, and stereochemistry of carbohydrates on MDP self-assembly. We then explore the utility of these glycosylated hydrogels as lectin delivery systems, showing affinity-based delayed release of lectins from hydrogels with specific carbohydrates. These works look to advance the use of MDPs as highly tunable drug delivery systems
Feeding Team Resilience: How Resilience is Cultivated
No full textTeam resilience is defined as a team’s ability to overcome challenges in a way that sustains performance over time. Gangs are teams that have consistently overcome adversity in a myriad of contexts and against a multitude of interventions. This Master’s thesis dissects how these teams build such strong bonds and, moreover, explores the types of behaviors gangs engage in to overcome challenges using semi-structured interviews with gang subject-matter-experts, including ex-gang members. The research team abductively analyzed the transcripts, finding four prominent themes across the data. Key findings include that gangs fulfill their members’ basic psychological needs, create an almost unbreakable team bond, a bi-dimensional hierarchy that takes advantage of people’s need for status, and lastly, that these groups strategically organize and delegate roles. Theoretical contributions, practical implications, and future directions are discussed
1.3 Public Infrastructure for Analyzing and Assessing Beyond Biocontainment Biotechnologies
Full text linkThis entreaty was created as part of The Spirit of Asilomar and the Future of Biotechnology summit (February 23-26, 2025) in Pacific Grove, CA.Arising from the Biotechnologies Beyond Conventional Containment (BBCC) theme of the 2025 Spirit of Asilomar Summit, this report proposes and outlines four integrated pillars of public infrastructure for testing organisms prior to possible release into open environments. The proposed infrastructure comprises physical, digital, human and stewardship elements. It would facilitate controlled and phased experiments in a manner similar to clinical trials, allowing for the development of predictive models regarding organism dispersal, persistence, gene flow, ecological impacts and more, across various environmental scenarios. Tiered access and governance mechanisms would ideally structure testing in relation to different risk categories, and promote transparent reporting of outcomes. Such infrastructure could transform how we assess the safety and value of releasing engineered living organisms into the environment, combining real-world data and computer models to improve informed choices based on solid evidence, responsible environmental care, and public input
Understanding and Suppressing Degradation in Mixed-Dimensional Halide Perovskites
Full text linkA new theory of 2D and 3D-2D perovskite degradation in humid air is demonstrated,
introducing the novel concept of an A-site cation – spacer cation – PbI2 – H2O quaternary phase
diagram. The degradation pathway of a given perovskite is determined by the path charted across
the phase diagram as water is introduced to the three-component (n=1) or four-component (n>1)
system, often moving through several multi-phase regions that include hydrate phases as nodes.
Because of the tendency of organic cations to escape the sample through solvation with the ambient
humid air, the effective composition of the system tends towards the PbI2 side of the phase diagram
and away from the organic side over time, while also increasing in H2O content. With this
degradation picture in mind, a new set of design principles can be conceptualized focusing on
suppressing the formation of hydrates and other “avoidable” degradation phases besides PbI2. This
new conceptual approach has been explored using the series of linear alkylammonium 2D
perovskite spacer cations from methylammonium to octylammonium, and the linear
alkyldiammonium spacer cations from butyldiammonium to decyldiammonium, both of which
form series of quasi-1D ribbon-like hydrate phases. The new conceptual framework can explain
the lower moisture stability of Dion-Jacobson phase 2D perovskites, and also explains the various
pathways seen for degradation of higher-n-value 2D and mixed 3D-2D perovskites in humid air. I
exploit the quantized ribbon length of the hydrate structure to engineer spacer cations so as not to
fit within the organic site of any ribbon-like hydrate, reducing the number of “avoidable” phases
in the phase diagram and as a result improving the moisture stability
Essays on Finance and the Real Economy
Full text linkThis dissertation contains three chapters. In the first chapter, I exploit variation in completion across planned bank branches to disentangle selection and treatment effects of branch entry on local economic growth. Areas where a bank planned to open a branch, but did not, exhibit higher growth than similar areas (reflecting a selection effect). However, locations where a bank opened a branch only slightly outgrow locations where a bank planned to open a branch but did not (treatment effect). Both effects are limited to the immediate geographic vicinity of proposed branches. These findings contrast with previous studies reporting positive treatment effects of branch entry and instead emphasize banks’ skill in selecting locations poised for growth.
The second chapter studies the impact of opening a stock exchange on a nation’s economic outcomes. Using a synthetic difference in differences approach, we estimate that opening the first exchange in a country increases national ten-year per capita GDP growth by more than 13 percentage points. This effect strengthens over time and is driven primarily by countries with strong governance. We evaluate the spatial distribution of this growth effect using satellite data. We estimate that opening an exchange increases ten-year growth in light emissions by 28 percentage points in the city where the exchange is located. The effect is geographically concentrated within 50 miles of the exchange, consistent with a causal interpretation of our results.
The third chapter examines whether mergers improve operational efficiencies, which remains unclear because constructing appropriate counterfactuals is difficult. Using synthetic controls, we generate bespoke (i.e., merger-specific) counterfactuals that more precisely reflect pre-merger trends and characteristics than traditional methods. We find that the average merger leads to small improvements in profitability, stemming from increased markups which outweigh declines in operational efficiency. Acquirers are roughly twice as likely to experience increases in markup and decreases in efficiency as to experience the opposite. These effects vary substantially across mergers and relate to firm characteristics in a way consistent with agency and market power theories