71446 research outputs found
Sort by
Resiliency, Distress, and Stress Coping in Adolescent and Young Adult (AYA) Cancer Survivors who Underwent the Bounce Back Program, a Stress Management and Resiliency Program, Across Gender and Race/Ethnicity
No full textAdolescent and young adult (AYA) cancer survivors are an overlooked population when it comes to mental health outcomes. Additionally, there are gender and race/ethnicity-based differences in levels of resiliency, distress, stress coping, and mind-body intervention engagement, however these have not been explicitly explored among AYA cancer survivors. This thesis aims to evaluate gender- and race/ethnicity-based differences in baseline, and pre- to post-intervention changes in resilience, depression, anxiety, and stress coping, as well as differences in engagement among participants of the Bounce Back program, a mind-body stress-management and resiliency program targeting the post-treatment phase among AYA cancer survivors. Independent samples t-tests were conducted to evaluate whether there were differences by gender and race/ethnicity with regard to mean baseline measures and mean number of sessions attended. Chi-square tests were performed to determine whether the completion status varied by gender and race/ethnicity. Linear regressions were performed to evaluate main effects of gender and race/ethnicity predicted on baseline and change scores of resilience, depression, anxiety, and stress coping, as well as to account for potential covariates. Baseline anxiety scores among females (M = 62.02, SD = 6.92) were significantly higher than among males (M = 54.42, SD = 10.37) (p = 0.01), even after adjusting for covariates. There were no differences between genders or races/ethnicities with respect to number of sessions attended or pre- to post-intervention changes in outcome measures, highlighting that the Bounce Back program is able to engage participants equitably and show similar improvements regardless of gender or race/ethnicity.Extension Studie
Biochemical Characterization of Interactions Between O-GlcNAc Transferase and Its Protein Binding Partners
No full textO-GlcNAc transferase (OGT) is an essential mammalian enzyme that is solely responsible for the post-translational modification of O-GlcNAcylation, the addition of N-acetyl-D-glucosamine (GlcNAc) onto serine and threonine sidechains on thousands of proteins in the nucleus and cytoplasm. OGT catalyzes another post-translational modification: the cleavage of host cell factor 1 (HCF-1), which it does via O-GlcNAcylation of a glutamate residue in one of six highly conserved proteolytic repeats. OGT also engages in a wide variety of protein complexes, facilitated primarily by its superhelical tetratricopeptide repeat (TPR) domain. Previous work from the lab has demonstrated that OGT’s O-GlcNAcylation and noncatalytic scaffolding functions are essential for cellular proliferation.
In this thesis, I build on previous notions of interactions between OGT and both its protein substrates and non-substrate interactors, defining features that drive these essential functions. Chapter 2 details a series of work performed in collaboration with other members of the Walker lab, studying the effects of mutations and truncations within the TPR domain on OGT’s functions and its ability to support cell proliferation; we show that truncation of OGT’s N-terminal TPRs impacts viability and all three functions, while preventing OGT homodimerization leads to increased cell growth. Additionally, in that chapter, I show that changes to protein-interacting features in the TPR lumen do not impact all of OGT’s substrates equally. In Chapter 3, I harness TR-FRET technology to develop a quantitative binding assay for OGT-substrate interactions. Using that assay, I identify a polypeptide derived from HCF-1’s proteolytic repeats, HCF3R, that has picomolar affinity and inhibits OGT in vitro and in cells. I then modify this probe to create a cellular OGT inhibitor that can be switched on and off by addition of two small molecules. In Chapter 4, I use a combination of biochemical and structural tools to interrogate the nature of the OGT-HCF3R interaction, finding that the medial TPR region (TPRs 6-8) contains features that are essential for the observed phenotypes. Taken together, this work demonstrates methods that can be used to further probe OGT’s functions, and insights by which new understanding of OGT’s protein-protein interactions can be derived.Chemical Biolog
Can Oyster Gardens Thrive in a Creek Condemned for Shellfish Harvesting? Cold Harbor Creek, Montross, Virginia
No full textCold Harbor Creek in Montross, VA, lies within the Chesapeake Bay watershed and is currently closed to shellfish harvesting. To examine whether Cold Harbor Creek is habitable for oysters, nine docks were recruited to support nine cages of oyster shells colonized with oyster spat. These populations were monitored over a period of eight months, and their growth was measured and evaluated. At the same time, water quality parameters were measured (dissolved oxygen, temperature, salinity, pH, and turbidity) immediately surrounding the oyster cages to determine whether the local conditions could impact the growth of the oysters. Approximately 75% of the cages maintained growing oyster populations, while the remaining cages showed limited survivorship. Water quality at most docks was generally similar to the conditions at the control sites, but decreased dissolved oxygen and pH were observed at the site with the lowest survivorship. The small population of oysters used in this project did not have a significant impact on local water quality. However, their growth indicates that water quality would not preclude a larger population of these organisms, which may have a greater effect on the Cold Harbor Creek water quality and could represent a biological remediation method in this condemned watershed.Extension Studie
Leveraging Functional Genomics to Elucidate Mechanisms of Targeted Protein Degradation
No full textTargeted protein degradation (TPD) is a rapidly developing field that hijacks the cell’s ubiquitin-proteasome system by using small molecules to induce proximity between a protein of interest and an E3 ubiquitin ligase, leading to target ubiquitination and subsequent proteasomal degradation. These TPD-inducing molecules, termed degraders, offer advantages over traditional occupancy-based inhibitors by operating through event-driven pharmacology, making this modality attractive as chemical biology tools and therapeutics. Molecular glues represent a unique class of degraders due to their monovalent nature and favorable drug-like properties, making them attractive both as therapeutics and as scaffolds for the development of heterobifunctional degraders. These molecules are also uniquely selective due to the high degree of protein and ligand cooperativity required for productive ternary complex formation between the neosubstrate and the E3 ubiquitin ligase. However, the rational development of glue degraders has been challenging, with most known examples identified through serendipitous discovery or high-throughput screening. Thus, strategies that enable the elucidation of their structural prerequisites, E3 ligase scaffold preferences, and potential resistance mechanisms will be important for accelerating the discovery and design of molecular glue degraders.
In this thesis, I present three projects that leverage functional genomics approaches, including CRISPR-suppressor screens, base editor scanning, and deep mutational scanning, to illuminate molecular principles governing TPD by molecular glue degraders. Genome editing technologies enable systematic amino acid-resolution mutagenesis in endogenous contexts, while expression library screens allow comprehensive evaluation of genetic variation in a controlled, unbiased manner to map structure–function relationships of E3 ligases. Together, these approaches reveal novel insights into mechanisms of action, resistance, and chemical-genetic convergence, which can be harnessed to guide rational degrader design.
In Chapter 2, I describe our efforts to map the landscape of resistance mutations in the neosubstrates RBM39 and GSPT1 in response to their respective molecular glue degraders. Through CRISPR–Cas9 tiling screens, we uncover canonical degron mutations that disrupt degrader binding, as well as distal site mutations that partially rescue degradation yet are sufficient for cell survival. Integration with evolutionary sequence conservation data reveals varying levels of sequence conservation across resistance sites in RBM39 and GSPT1, suggesting that structural and functional requirements of protein regions constrain the accessible mutational space and thereby drive divergent mutational outcomes across targets.
In Chapter 3, I focus on elucidating the mechanism of action of the molecular glue degrader UM171, which induces the degradation of the LSD1–CoREST corepressor complex via the Cullin3-RING E3 ligase substrate receptor, KBTBD4. By integrating proteomics, functional genomics, biochemical analysis, and cryo-electron microscopy (cryo-EM), we show that HDAC1/2, the direct target of UM171, bridges CoREST to KBTBD4, and that CoREST degradation requires a dual-glue mechanism involving the metabolite inositol hexakisphosphate (InsP6). Base editor scanning further validates critical residues in the ternary complex interface, showcasing the value of precise genetic perturbations in probing requirements for degrader activity.
In Chapter 4, I investigate KBTBD4 mutations found in medulloblastoma, demonstrating that these mutants are gain-of-function (GOF) and mimic UM171 by promoting neomorphic degradation of CoREST through HDAC1/2 engagement. Using deep mutational scanning, we chart the mutational landscape of the KBTBD4 cancer hotspot, revealing a strong preference for insertion mutations in driving neomorphic E3-substrate engagement. Unexpectedly, structural analyses demonstrate a convergent shape complementarity phenomenon between these cancer mutations and UM171, where genetic mutations and chemical matter structurally and functionally mimic each other. We also demonstrate that HDAC1/2 inhibitors can disrupt the mutant ternary complex and suppress proliferation in patient-derived KBTBD4-mutant medulloblastoma cells, revealing a therapeutically actionable vulnerability.
Collectively, these studies illustrate the power of massively parallel genetic screening in uncovering mechanistic principles of molecular glue degraders, identifying resistance vulnerabilities, and guiding degrader design, prompting their use as valuable tools in the TPD space. Importantly, we describe a phenomenon of chemical-genetic convergence where genetic mutations and small molecules phenocopy each other structurally and functionally, highlighting a framework where genetic perturbations can be used to aid the rational design of induced proximity therapeutics. More broadly, this work underscores the potential of functional genomics to accelerate the discovery of next-generation degraders and proximity-based therapeutics.Chemistry and Chemical Biolog
Service Design Strategies for On-Demand Operations
No full textAs the service industry accelerates, companies face significant operational challenges in designing service strategies that optimize efficiency while enhancing customer and worker experiences. This dissertation explores key operational decisions in last-mile delivery and on-demand fulfillment, focusing on how service design choices impact customers' and service providers' performance outcomes. Through field experiments and observational data, I examine the effects of communication strategies, customer location heterogeneity, and the design of monetary incentives on the efficiency of on-demand operations.
The first chapter investigates how the choice of communication channel affects delivery success rates. In collaboration with a logistics platform that uses decentralized residential inventory storage for same-day fulfillment, I study the outcomes of a field experiment comparing email-based (informational) and WhatsApp-based (interactive) delivery notifications. Contrary to expectations, although the interactive channel increases engagement, it also leads to a higher rate of failed deliveries. This effect is driven by customers attempting to coordinate the delivery times and locations, despite the channel not being monitored in real time. These findings highlight the importance of channel norms when designing services and using customer communication to improve operational efficiency.
The second chapter examines how customer proximity to fulfillment centers (FCs) influences purchasing behavior and service costs. Using data from an on-demand grocery delivery company in Latin America, I show that small differences in delivery promise times significantly impact order frequency and that--beyond the promise--actual delivery time matters as well: customers farther from FCs experience longer delivery times and purchase less frequently. This spatial effect can lead to a "catchment area shrinking effect", where FCs predominantly serve nearby customers while losing engagement from those further away. The impact of proximity in an online-only service suggests that managers should incorporate distance into service design decisions--such as their network design, pricing, and delivery promise times.
The third chapter focuses on the supply side of on-demand services, analyzing how incentive design can increase gig workers’ retention. Using data from a platform that provides B2B delivery services in which drivers choose their daily routes, I study how the change in incentive structure from a bonus that compensates drivers on taking an extra route (extensive margin) to one that compensates them for the number of deliveries on the extra route (intensive margin) increases both the effectiveness of the incentive and long-term driver retention. I discuss monetary and non-monetary mechanisms that may drive the persistent effects of short-lived incentives, such as goal setting and learning. These results highlight the value of incentive design, a short-term lever, for worker retention in the long run.
Together, these studies contribute to the field of service operations management by offering empirical insights into service design strategies for on-demand fulfillment. The findings inform platform-based businesses on how to improve customer interactions, manage geographic service distribution, and enhance worker retention--ultimately improving the operational efficiency of on-demand service models.Business Administratio
Audio Processing on Field-Programmable Gate Arrays
No full textThis document outlines the methods employed from conception to implementation and testing of a hardware-based digital audio processing system. The eight-month project was divided into phases of problem definition, system design, implementation via register-transfer level (RTL) code, and finally measurement and analysis of the design. The system was designed with live performance applications in mind and aims to achieve technical specifications that enforce high reconfigurability and signal quality, minimal latency, and low utilization of hardware resources. These technical specifications define a high-performance, low-cost system that acts as a proof-of-concept system that, with further development, has the
potential to outperform the state-of-the-art with regard to both adaptability and pricing without sacrificing latency or signal quality.Electrical Engineering S
Disease-Modifying Antirheumatic Drugs, Rheumatoid Arthritis-Associated Lung Disease, and Serious Infection: Risks and Associations
No full textAbstract for paper 1:
Objective: To investigate the association of disease-modifying antirheumatic drugs (DMARDs) and risk of incident interstitial lung disease (ILD) among patients with rheumatoid arthritis (RA) using a systematic literature review and meta-analysis.
Methods: We performed a systematic literature review and meta-analysis of studies examining the association of DMARDs with incident RA-ILD. PubMed, Embase, Web of Science, and Cochrane Library were searched from inception to November 2023 for randomized controlled trials (RCTs), observational studies, and post-marketing surveillance studies that investigated adults with RA and compared DMARDs of interest with placebo, no DMARDs, or other DMARDs. The outcome was incident ILD. We summarized the literature on DMARDs and incident RA-ILD risk. Among studies with sufficient quality, we performed meta-analyses to obtain odds ratios (OR) and 95% confidence intervals (95%CI) using the Mantel-Haenszel method.
Results: Among 3,612 studies, we identified a total of 40 papers that encompassed 486,465 patients with RA and 3,928 incident ILD outcomes that were included in the final systematic review and meta-analysis. Among the studies, 24 were RCTs, 4 were prospective cohort studies, 9 were retrospective cohort studies, 2 were case-control studies, and 1 was a post-marketing surveillance study. The pooled analysis from RCTs revealed no statistically significant difference in the odds of ILD development for any specific DMARD across all comparisons examined. The largest identified RCT (Oral Surveillance trial) of tofacitinib (n=2,911) vs. tumor necrosis factor inhibitor (TNFi, n=1,451) found no relationship with incident ILD (OR 0.94, 95%CI 0.52 to 1.69, p=0.828). In 7 observational studies, the use of methotrexate (MTX) yielded a pooled OR for ILD of 0.49 (95%CI 0.32 to 0.76, p.001) compared to those not using MTX. In a single observational study, tofacitinib users had an OR for ILD of 0.36 (95%CI 0.15 to 0.87, p=0.024) compared to TNFi users.
Conclusion: Observational data suggest no increased risk of MTX and tofacitinib for incident RA-ILD. However, these studies may be susceptible to bias, and no specific DMARD showed associations with incident RA-ILD in RCTs. Further well-designed prospective studies are warranted for definitive conclusions on the potential relationship between DMARDs and RA-ILD risk.
Abstract for paper 2:
Objective: To investigate the association between rheumatoid arthritis-associated lung disease (RA-LD) and serious infection risk.
Methods: We conducted a retrospective cohort study using the MGB Biobank (Boston, Massachusetts), comparing RA-LD to RA patients without lung disease (RA-no LD), matched by age, sex, and RA duration. RA-LD cases were verified by medical record review and chest imaging for clinically apparent RA-associated interstitial lung disease (RA-ILD) and/or RA-associated bronchiectasis (RA-BR). The primary outcome was serious infection. Incidence rates and propensity score-adjusted subdistribution hazard ratios (sdHR) were calculated using Fine and Gray models to account for competing risk of death.
Results: Among 221 RA-LD cases (151 RA-ILD and 70 RA-BR) and 980 RA-no LD comparators, RA-LD had a significantly higher serious infection risk compared to RA-no LD comparators (55.8 vs. 25.8 per 1,000 person-years, sdHR 1.60, 95%CI 1.20-2.12). The increased risk remained significant for RA-ILD cases (sdHR 1.79, 95%CI 1.33-2.41), but not for RA-BR (sdHR 1.19, 95%CI 0.72-1.97). Anatomic sites of infection that were more common in RA-LD included pulmonary, skin and soft tissue, and ear, nose and throat; RA-LD was associated with various pathogen types: virus, bacteria, fungus, and mycobacteria. Specific pathogens with higher frequency in RA-LD cases, particularly among RA-BR, included influenza virus, respiratory syncytial virus, Staphylococcus, Pseudomonas, and nontuberculous mycobacteria.
Conclusion: RA-LD, particularly RA-ILD, is associated with a significant increased risk of serious infection across anatomic sites and diverse pathogen types. RA-BR is associated with increased pulmonary infections. Tailored infection prevention strategies, including vaccination, are essential for this vulnerable population.Graduate Educatio
From Policy to Practice: Physician Decision Making, Patient Outcomes, and Policy Evaluation in Global Tuberculosis Control
No full textTuberculosis (TB) remains one of the world’s leading causes of infectious disease death, disproportionately affecting individuals living in low- and middle-income countries. While biomedical advances have improved diagnosis and treatment, persistent gaps in care reflect broader system-level, clinical, and policy challenges. This dissertation explores global TB control through three interconnected perspectives: physician decision-making, patient outcomes, and health policy evaluation, with a focus on real-world implementation in resource-constrained settings.
Chapter 2 examines how clinicians decide to initiate TB treatment for individuals being evaluated for TB when bacteriological test results are negative. Using a systematic review and individual patient data meta-analysis from over 15,000 individuals across 13 countries, I identify factors—such as symptom profile, HIV status, and prior TB history—that influence treatment decisions. The findings highlight the complexities of diagnostic decision-making in resource-limited settings.
Chapter 3 investigates long-term mortality among individuals diagnosed with TB in Brazil. Leveraging linked administrative data, I construct a retrospective cohort of over 700,000 individuals diagnosed with TB and compare their mortality risk to that of the general population over time. Results show elevated mortality persisting for many years after treatment, highlighting the importance of continued care and monitoring beyond the treatment period.
Chapter 4 evaluates the potential health and economic impact of targeted active case finding (ACF) in Blantyre, Malawi, guided by community-level Mycobacterium tuberculosis immunoreactivity surveys. Using a simulation model, I compare survey-informed geographical targeting to untargeted ACF and passive case finding. Findings suggest that targeted ACF guided by immunoreactivity survey results can improve efficiency, particularly in settings with a high prevalence of undiagnosed TB and high HIV burden.
Together, these chapters illustrate that advancing global TB control requires more than technical fixes—it demands engagement with clinical uncertainty, long-term patient needs, and the complexities of health system design. This dissertation highlights the importance of aligning diagnostic approaches, treatment pathways, and policy responses with both patient realities and system-level constraints, contributing evidence to support more responsive and equitable TB programs.Population Health Science
Essays on the Impact and Effectiveness of Share Repurchase Regulations
No full textThis dissertation contributes evidence on the effectiveness of different share repurchase regulations. Share repurchases, also known as stock buybacks, occur when a firm buys back its own stock, usually on the public market. Over the last three decades, repurchases have emerged as an important, yet controversial, form of shareholder payouts. Compared to dividends, repurchases give managers greater flexibility in the timing and magnitude of their firm's payouts. This allows managers to better manage temporary increases in cash and changes in capital structure, in addition to signaling undervaluation. With that said, critics are concerned that managers may misuse repurchases. Managers may have incentives to forego positive NPV long-term investments to increase stock prices in the short-term, coming at the cost of long-term firm value.
One option to curtail the potential harmful impacts of repurchases is to increase transparency surrounding them. In solo-authored work, the first paper (Chapter 2) considers whether additional share repurchase legalization can reduce the frequency of value-decreasing repurchases. The SEC adopted its Share Repurchase Disclosure Modernization Rule (the ``modernized rule") in 2023. Relative to the previous disclosure rules, the modernized rule required additional disaggregation of quarterly repurchases to better show managerial repurchase intent. The modernized rule became effective in October 2023, but was vacated in December 2023. Due to different quarter start dates, the modernized rule was effective for some firms for nearly two months; other firms were never treated. Using this quasi-exogenous variation, monthly repurchase data and a staggered DiD design, my main tests find that the modernized rule reduces share repurchases amounts, primarily by reducing the number of repurchasing firms. In cross-sectional tests, efficient (``value maximizing") repurchases decline significantly, while opportunistic (``value-decreasing") repurchases do not. The decline in efficient repurchases is consistent with the legal, liquidity and proprietary costs of disclosure outweighing the benefits of reduced information asymmetry. Meanwhile, the limited reduction in opportunistic repurchases suggests that additional disclosure has little to no impact on the manager’s private cost of repurchasing too many shares. Results using event study returns and repurchase prices corroborate these effects. Overall, the findings are consistent with additional repurchase disclosure having real effects on repurchases. It reduces repurchases; however, efficient repurchases decline and opportunistic repurchases are unaffected.
Other critics of repurchases have argued for significant restrictions or even banning repurchases altogether. In co-authored work with Charles Wang, the second paper (chapter 3) evaluates these claims by re-examining the \textit{purported negative impact} of share repurchase legalization on corporate investment using staggered legal changes across 17 countries from the 1980s to 2000s. Considering \textit{all} public firms instead of just \textit{repurchasing} firms, we document increases in investment and firm performance. Cross-sectional tests are consistent with capital from repurchasing firms flowing to smaller, younger, higher-growth, and more cash-needy companies. The third paper (Chapter 4) challenges the robustness of Wang, Yin and Yu's (2021) direct finding that repurchasing firms reduce their investments after repurchase legalization. Employing more robust econometric methods and additional analyses, we find no consistent evidence that legalizing buybacks reduces investment among repurchasing firms. Thus, legalization does not appear to be harmful to repurchasing firms' investments either. These results suggest that legalizing repurchases may facilitate efficient capital reallocation rather than curtail aggregate investment. They contribute to the ongoing debate about the economic effects of share repurchases and have important implications for corporate finance literature and public policy. Policymakers should exercise caution when considering restrictions on share repurchases, as such policies may inadvertently hinder efficient capital allocation.Business Administratio
Risk Factors Associated with Progression and Future Outcomes in Patients Undergoing Coronary Angiography
No full textCoronary artery disease (CAD) is a complex disease influenced by multifaceted risk factors, including genetics and novel biomarkers. Despite advancements in managing traditional cardiovascular risk factors, CAD continues to pose major risks of mortality and morbidity globally. Many patients sustain adverse outcomes and complications of CAD despite optimal medical therapy and excellent control of these risk factors. This persistent risk underscores the need for a deeper understanding of residual risk to better stratify patients with CAD, highlighting the importance of exploring novel biomarkers that could better predict the risk of outcomes, especially in a secondary prevention context.
Given the high heritability of CAD, integrating genomic information plays an important role in patient risk stratification. Genetic factors, including germline variations–both monogenic, such as Familial Hypercholesterolemia (FH), and polygenic liabilities, as represented by a polygenic risk score (PRS)– and somatic variants, such as clonal hematopoiesis of indeterminate potential (CHIP) have been robustly linked to the incidence of CAD in large cohort studies. However, the role of these genomic drivers in a secondary prevention paradigm, particularly after coronary angiography, requires further extensive exploration.
Lipoprotein(a) [Lp(a)] is another important marker of residual risk associated with major cardiovascular events. This low-density lipoprotein, whose circulating concentrations in the plasma are largely genetically determined, is elevated in approximately 20% of the world’s population. Every 50 nmol/L increase in Lp(a) is associated with an 11% increased risk of incident atherosclerotic cardiovascular disease. Currently, American and European guidelines recommend Lp(a) testing at least once in a lifetime for adults, regardless of traditional cardiovascular risk factors. Despite its recognized role, the impact of elevated Lp(a) on long-term outcomes following coronary angiography remains underexplored, necessitating further studies into its implications for secondary prevention.
To investigate these aspects, we studied patients undergoing coronary angiography, the gold standard for diagnosing and evaluating coronary phenotypes and severity. In this cohort of individuals with CAD or a high likelihood for the disease, we aimed to understand whether (1) genomic predisposition to CAD and (2) the novel lipid biomarker, Lipoprotein(a), correlate with specific angiographic characteristics and track progression and future outcomes. We hypothesized that these factors are associated with more severe coronary phenotypes and increased risks of adverse outcomes, either clinically or angiographically, following initial coronary angiography.Graduate Educatio