22689 research outputs found
Sort by
Assessing the mental health of multiracial and multiethnic adults in the United States
No full textAbstract
Objectives
Multiracial and multiethnic people are part of the fastest growing demographic in the United States, yet little is known about their mental health. This multi-study dissertation explores mental health among multiracial and multiethnic adults, and the impact of intersectional identities, environmental factors, potentially traumatic experiences, prejudice events, social and interpersonal support, social inclusion, and strength in ethnic identity.
Methods
A cross-sectional survey instrument was developed that assessed for symptoms of depression, anxiety, post-traumatic stress disorder (PTSD), and suicidal behaviors; exposure to potentially traumatic events and prejudice events; estimated perceived stress, strength in ethnic identity, and perceived social support; and collected demographics. The initial round of surveys were collected from February-June 2022, sampling through social media platforms and listservs for multiracial and multiethnic communities, with a final sample of 347 (A). A second round of surveys were collected October 25-December 6, 2022 using paid panels facilitated by Qualtrics, with a final sample of 1012 (B). To participate in either study, respondents had to be an adult (18+) who lived in or were from the United States and self-identify as multiracial and/or multiethnic. Bivariate and multivariate logistic regression models were conducted on each sample separately.
A subset of participants (n=17) from the first round of data collection who endorsed one or more mental health concerns in the survey participated in semi-structured interviews December 2022-January 2023. These interviews explored sentiments related to self-disclosure of racial and ethnic identity; attitudes, experiences, and practices related to mental health; and policy and programmatic priorities to support mental health and resilience of multiracial and multiethnic communities. This series of studies uses participatory action research methods and is led by a member of the multiracial and multiethnic community, includes multiracial and multiethnic adults in research design, instrument development, data analysis, interpretation of findings, and dissemination.
Results
Participants in each survey sample endorsed symptoms for depression (A: 40.6%, B: 42.1%), anxiety (A: 41.5%, B: 40.5%), PTSD (A: 32.9%, B: 40.4%), and suicidal behaviors (A: 21.1%, B: 25.4%) higher than overall estimates from nationally representative surveys. Multivariate analyses suggest a gradient effect of more lifetime exposure to potentially traumatic experiences (OR=1.13, p<.001), discrimination (OR=1.12, p=.003), and microaggressions (OR=1.02, p=.002) on the odds of having one or more mental health outcome. High perceived social support (OR=.14, p<.001) and affirmation of ethnic identity (OR=.91, p=.001) were inversely associated with adverse mental health.
Participants in the qualitative interviews identified self-disclosure of race and ethnicity as a unique stressor for multiracial and multiethnic populations due to inconsistency in question structure and available categories, perceived mismatch of racial and ethnic identity and phenotype, and exposure to prejudice. Social norms, constructs, and movements impacted the racial and ethnic categories a multiracial and/or multiethnic person would provide to external parties. The ability to select multiple categories and/or respond to an open answer question were connected to positive sentiments. Multiracial and multiethnic adults described navigating bias and stigma from cultures, communities, and providers when seeking mental wellness. Respondents demonstrated resilience and strength, and emphasized the importance of secure identity formation, impact of group specific processes, reducing barriers to mental health care, and need for culturally responsive care.
Conclusions
This series of studies highlights the complex mental health needs of multiracial and multiethnic adults in the United States requiring future research and public health intervention. Multiracial adults have a high prevalence of mental health conditions, consistent with prior research. Intersectionality, prejudice events, strength in ethnic identity, and social support impact their mental health
RNA FOLDING AND THE FITNESS LANDSCAPE OF THE glmS RIBOZYME
No full textThe discovery of catalytic RNAs suggests that life emerged from an RNA molecule. However, the principle of how an RNA sequence evolved to acquire stabilizing tertiary motifs for optimizing the function is beginning to be understood. By mapping an activity landscape of numerous variants, one can decipher the connection between sequence and function and deduce the evolutionary rules that favor a particular structural motif. The glmS ribozyme is part of a negative feedback loop that maintains cellular levels of glucosamine-6-phosphate (GlcN6P). The self-cleavage of the ribozyme in complex with GlcN6P leads to the turnover of the glmS mRNA. The cleavage of RNA depends on proper folding of the ribozyme sequences, which provides a convenient readout of the sequence-function relationship. In addition to developing a medium-scale screening approach to generate a hierarchy of mutation tolerance for structural domains, I combined the self-cleavage assay with high-throughput sequencing to estimate ribozyme activity and project the activity profile onto the secondary and tertiary structures to rationalize the mutational effects. By surveying 456 single mutants, I observed that many deleterious mutations are clustered in the phylogenetically conserved catalytic core. Other detrimental base substitutions lie in the internal loop IL4, which forms tertiary interaction with the core helices. The thermally metastable conformer caused by the IL4 mutation suggests that sequence conservation in the core-periphery contact enhances the ribozyme function by preventing RNA misfolding. The investigation of 14205 pairwise epistatic interactions in a condition that stabilizes the folded RNA showed that mutations in different structural domains often create negative epistasis. In contrast, mutation pairs in the same motif show positive epistatic interactions. Moreover, the analysis of epistasis change among different folding conditions demonstrated that folding cooperativity is stronger in the crowding condition. On some rare occasions, one harmless mutation compensates for the other deleterious mutation in the same peripheral motif. This compensatory interaction indicates an alternative folding pathway that may exist during RNA evolution. This study has provided indications of molecular evolution by scrutinizing the link between sequence and cleavage activity, allowing a prospective understanding of molecular decisions for conserving tertiary motifs that promote RNA functionality
THE ROLE OF DEEP, REDUCING CRUSTAL AND UPPER MANTLE FLUIDS IN THE MOBILITY OF CHROMIUM AND THE EVOLUTION OF HYDROCARBONS
No full textFluids have played important roles at deep crustal and upper mantle conditions, especially in subduction zones. Fluid-rock interactions are vital for the cycles of inorganic and organic species in the deep Earth. The mobility of species should be in principle dependent on temperature, pressure, chloride concentration, and oxidation state. However, most studies of subduction zone fluids have focused on oxidizing conditions. Here, it is proposed that metals, especially chromium, and carbon have drastically different behavior in aqueous fluids at lower oxidation states compared to higher oxidation states. For chromium, it is suggested to be mobile as Cr(II) in aqueous solutions at high temperatures, rather than Cr(III) under ambient conditions. In the case of aqueous carbon, a wide range of species with oxidation states from C(IV) to C(-IV) can exist at elevated temperatures and pressures. Furthermore, there is the potential for the development of separate C-rich fluids immiscible with an aqueous phase.
This dissertation is motivated by studying the chemical evolution of deep, reducing crustal and upper mantle fluids, focusing on the mobility of inorganic Cr species and the formation of immiscible hydrocarbon fluids. This is done through three specific objectives. In the first study, a new Cr(II) complex CrCl(OH)0 was regressed in the Deep Earth Water model using experimental solubility data for Cr2O3. A high solubility was predicted in saline, peridotitic diamond-forming fluid, which implicated that diamond formation may involve simultaneous oxidation of Cr(II) to Cr(III) and reduction of C(IV) to C(0). In the second study, in situ experiments were performed using a diamond anvil cell to investigate the effect of pressure on the formation of immiscible hydrocarbon fluids at subduction zone conditions. Hydrocarbon fluids could perhaps play an important role in the deep carbon cycle. In the third study, an irreversible chemical mass transfer model was carried out to simulate the formation of immiscible CH4-rich fluids in the alteration of ophicarbonates in the Lanzo Massif, previously described by Vitale Brovarone et al. (2017). This work provides insight into the potential importance of reducing conditions in subduction zone systems
Simulation-Based Optimization with Constrained SPSA for Water Distribution Networks on Military Installations
No full textThe purpose of this paper is to combine simulation-based optimization and simultaneous perturbation stochastic approximation (SPSA) to create an effective model of a water distribution network and return the optimal diameters for the system. This paper particularly focuses on a distribution network for a military installation. Using a water network simulation that includes random processes to model real world variability, we minimize the monetary cost and amount of the population that receives an inadequate amount of water. We use sequential quadratic programming and projection constraints to add bounds to our model. We conclude by showing that in two case studies, our model using simulation-based optimization performs better than the previously established pipe diameters in the networks
Uncrewed Aerial Vehicle Fruit Picking with Perceptual Imitation Learning Trajectory Generation
No full textThis thesis studies the problem of Uncrewed Aerial Vehicle (UAV) path planning and manipulation in unmapped environments. This thesis the specific task of orange picking with a quadrotor UAV. Robotic fruit harvesting is a fitting example problem to tackle in this research, as there is a worldwide need for improved agricultural technologies.
This task is difficult because it requires comprehending and navigating a complex, unknown environment.
To accomplish this task, we present a novel visual servoing controller which fuses information from onboard camera images with odometry data. This was used to calculate the relative position of an orange and a safe approach angle. By following a series of reference trajectories to the computed goal location, the system was able to grasp an orange autonomously and remove it from the tree.
This visual servoing method has several inherent limitations. It cannot search for an occluded orange or handle any paths that remove the orange from its view. To improve upon this approach, and correct these shortcomings, we develop a novel neural network architecture to perform the same task using a learned implicit visual encoding.
In the next section, we present the design of a simulation of this same orange picking task, and a Model Predictive Control (MPC) method for computing optimal trajectories within it. We trained the neural network to imitate the MPC expert, validating the network structure and cost function.
In the subsequent chapter, we trained the same architecture on a dataset derived from the visual servoing controller. These experiments led to useful innovations in the neural network architecture, but even with these efforts, no network was able to vastly improve on the baseline data.
In the final chapter, we discuss the relative strengths and weaknesses of these algorithms. Each has areas where it exceeds the others, and we propose new avenues of research to improve them all
Multi-layer and heterogeneous integrated photonics
No full textIntegrated photonic circuits are the optical analog to the ubiquitous electronic integrated circuits that are found in many everyday items. Much like their electronic analog, photonic integrated circuits have enabled the miniaturization of many complex optical systems and result in smaller size, less power consumption, reduced cost, and larger bandwidths. As the field of integrated photonics has matured, the complexity, desired capabilities, and library of material platforms have expanded. This expanded library of materials has allowed certain materials to be used for specific applications based on the strengths of their intrinsic properties. Such an approach inevitably leads to design trade-offs because any complex optical system will require optimal performance in a variety of optical properties.
In this thesis, I will explore heterogeneous design via multi-layer integration using mature integrated photonic material platforms. I will demonstrate that a heterogeneous platform realized through multi-layer integration enables linear and nonlinear performance that cannot be achieved using a single-layer approach through various nonlinear optical demonstrations such as high-speed optical parametric gain, ring resonator enhanced four-wave mixing, and photon-pair generation via spontaneous four-wave mixing. In the second part of this thesis, I will add to the current library of integrated photonic material platforms through the investigation of a new material platform, niobium-tantalum oxide, for integrated photonics. I will demonstrate the utility of this material for linear and nonlinear optical applications and show that it is an ideal material for heterogeneous multi-layer integration
PHENOTYPIC CHARACTERIZATION OF NOVEL HER2 MISSENSE MUTANTS AND INVESTIGATING THE ROLE OF INFLAMMATORY SIGNALING IN ER+ ADVANCED BREAST CANCER
No full textEndocrine resistance is the major challenge in treating patients with advanced estrogen receptor-positive (ER+) breast cancer. The gaining of HER2-activating mutations is one of the mechanisms that confer endocrine resistance (Nayar et al., 2018). Knowledge about the phenotype of HER2 alterations in ER+ breast cancer remains incomplete, mainly due to the low mutation frequency and limited resources from clinical databases. Here, in the first project, we are applying deep mutational scanning (DMS) to comprehensively characterize the phenotype of all HER2 missense mutations in ER+ breast cancer. We have optimized PCR conditions to efficiently amplify HER2 open reading frame (ORF) from the genomic DNA (gDNA) for the final DMS screens. Meanwhile, we have selected and conducted functional validation for 20 putative activating HER2 mutants and identified 8 activating mutations that confer resistance to ER-targeted therapies and with varying responses to CDK4/6 inhibitors but are sensitive to the combination treatment of fulvestrant and neratinib, an irreversible tyrosine kinase inhibitor. In the second project, we focused on the inflammatory signaling in ER+ HER2mut breast cancer. We report the ERK-dependent upregulation of stimulator of interferon genes (STING) and the hyperactivation of its downstream signaling, with enhanced tumor immunogenicity in terms of increased cytokine production, PD-L1, and MHC-1 expression. We also found MYC is downregulated in T47D cells with activating HER2 mutations, which may also contribute to the augmented cytokine production, and ER appears to modulate MYC level in a ligand-induced manner. This study shows advanced ER+ breast cancer with targeted therapy resistance may have enhanced tumor immunogenicity and potential response to immune checkpoint blockade therapy
Can Tech Be Good for Health?: The Intersection of Responsible Investing, Digital Products, and Social Impact
No full textBackground: Technology has an increasing influence over the lifestyle behaviors that drive the rising incidence of chronic disease. This creates an opportunity for an untapped stakeholder, the private sector, to use their outsized influence to meaningfully contribute to improving public health. This dissertation explores the current landscape among private sector technology companies and impact-driven investors, internally and externally, to create evidence-informed recommendations for how these stakeholders can change their organizational behavior to be health-positive, which refers to intentionally designing products, programs, or services to have a positive influence on the health of consumers.
Aims: 1) Understand the impact landscape within the tech sector and determine the role for health-positive strategy 2) Discern the role of health in corporate reporting today and how to create conditions for accountability 3) Demonstrate how to apply the insights from internal and external perspectives for feasible health-positive product development.
Methods: Aim 1 will be addressed through semi-structured stakeholder interviews with 19 participants from consumer tech companies and impact-driven investors. Thematic analysis will be used to identify dominant themes. Aim 2 will use a Natural Language Processing (NLP) algorithm and thematic analysis to pull patterns from corporate proxy statements. Aim 3 will encompass a multiple-case study design.
Results: In Aim 1, interviews revealed the current state of health in impact strategy in tech, motivators and barriers for its integration and measurement, and the process for developing a health-positive product strategy. In Aim 2, content analysis of proxy statements determined health topics and initiatives in corporate reporting in the tech sector, and their relative prevalence. Lastly, Aim 3’s case study analysis concluded that there are existing levers through which companies can integrate health-positive thinking, particularly through product design, policy, and their implementation.
Conclusion: Business priorities and positioning were seen as the strongest indicator of a company’s likelihood to pursue health impact through products. Still, a lack of reliable, auditable metrics is hamstringing adoption. COVID-19 has emerged as an inflection point, surfacing a previously unseen lever to drive material business impact, and prompting action. Aim 1 and 3 produced recommended processes for integrating health into product strategy and the first steps towards developing auditable metrics. Aim 2 showed health impact is not absent in the private sector, but there is no unified language allowing for performance evaluation and accountability. Barriers lie in a persistent tension between product- vs. company-specific frameworks, corporate resistance to first-mover disadvantage, and consistent delay in tech policy and regulation. By establishing transparency, best practices, and market recognition, corporate health impact through the product environment can become an exciting, opportune intervention point for public health
MACHINE LEARNING BASED PREDICTION OF PANCREATIC NEOPLASIA IN HIGH-RISK INDIVIDUALS USING IMAGING BASED TEXTURE FEATURES
No full textPancreatic ductal adenocarcinoma (PDAC) is growing to be one of the leading cancers
across the world. Patients with PDAC have very low chances of survival. Surveillance
of high-risk individuals (HRI) for PDAC due to family history and/or genetic mutation
can lead to the early detection of PDAC improving their chances of survival. Precursors
to PDAC such as pancreatic intraepithelial neoplasia (PanIN) are usually microscopic
and hard to detect by traditional surveillance methods such as endoscopic ultrasound
(EUS) or magnetic resonance imaging (MRI). Precursors to PDAC are also associated
with increased fibrosis, the presence of higher levels of fat in the pancreas and the
destruction of acinar cells in the pancreas parenchyma which induce texture changes
in most pancreas tissue ultrasound imaging methods. These changes are usually
monitored subjectively by an endosonographer during a procedure. Quantifying these
echotexture changes in commonly used imaging-based diagnostic tools affords clinicians
the opportunity to detect these microscopic changes early on. Image texture-based
Machine Learning (ML) models could possibly detect the occurrence of PDAC in HRI.
The aim of this study is to analyze EUS images of HRIs to use image-based texture
features to predict final diagnostic pathology (Low-grade dysplasia [LGD] against
High-grade dysplasia [HGD] or PDAC).
We selected 41 HRI who had surveillance and surgical treatment for suspected
neoplasms within a multidisciplinary care program at an academic center. The final
pathological diagnosis was established as LGD or HGD or PDAC pre-surgery and
was confirming post-surgery using histopathological analysis of pancreatic tissues. To accomplish the aim, good quality EUS images are chosen by a team of endosonographers
within 1.5 years of surgery based on the images that capture the anatomic location of
surgery. These images are then annotated to identify pancreatic parenchyma using
regions of interest (ROIs) drawn over them. These manually annotated ROIs exclude
artifacts, cysts, ducts, vessels, lesions, and solid masses. Second-order texture metrics
such as Gray level matrices are computed to quantify texture features observed across
ROIs. 81 computed texture features that quantify the heterogeneity observed across
the images are then analyzed using an XGBoost model to classify the images into
two classes - LGD or HGD/PDAC. Cross-validation based model training and testing
are used to find the best classifier. The top six texture features contributing to the
predictive value of the classifier are used to compare histopathological features such
as fat, collagen, and PanINs associated with PDAC to understand the correlation of
texture features to anatomic changes.
EUS images from HRI had HGD/PDAC (44.5%) and LGD (55.5%) respectively
from a total of 726 images used to classify into the two groups. The 81 texture
feature model had a sensitivity of 71% [60-80%], specificity of 82% [72-89%], and
accuracy of 77% [69-83%] and an AUC of 0.75 [0.67-0.83]. The top six features
with high predictive power according to the classifier namely - Gray Level Run
Length Matrix (GLRLM) Run Variance, Gray Level Co- occurrence Matrix (GLCM)
Cluster Shade, GLCM Difference Average, GLCM Cluster Prominence, Gray Level
Dependence Matrix (GLDM) Dependence Variance, and GLCM Contrast are then
used to understand the correlation to histopathological features - fat, collagen, and
PanIN. A possible correlation is observed but a much larger dataset is necessary to
verify this correlation. Further testing and validation of these models can help develop
comprehensive surveillance of HRIs and understand the process of PDAC development
and improve effective therapy strategy
Composition Optimization of mRNA Lipid Nanoparticles to Modulate Immune Activation Profile and Potentiate Anticancer Immunity
No full textLipid nanoparticles (LNPs) have emerged as an effective platform for delivering mRNA encoding tumor antigens in cancer immunotherapy. Recent research has focused on optimizing LNPs to improve transfection and maturation of antigen-presenting cells, modulate Toll-like receptor (TLR)-mediated adjuvant activity, and enhance CD8+ T cell response and antitumor efficacy. This study utilized a multi-step screening method to optimize the helper lipid type and component ratios in an mRNA LNP library for efficient delivery of antigen-encoding mRNA and modulating the immune activation profile, specifically the balance between Th1 and Th2 responses. The screening process first involved assessing dendritic cell maturation and antigen presentation in vitro, followed by evaluating immune activation and tumor growth suppression in vivo. LNP formulations with potent antitumor efficacy were identified through this method, particularly in combination with immune checkpoint blockade therapy. Our findings also revealed that LNPs with dual immunostimulation activity, eliciting both Type-1 T helper (Th1) and Th2 responses, produced the most potent antitumor efficacy by triggering a coordinated attack by dual T cells and NK cells. These results highlight the crucial role of LNP composition, in addition to the payload, in tailoring immune responses and offer new opportunities to optimize RNA-based treatments for cancer and infectious diseases