22689 research outputs found
Sort by
TIME AND CAUSALITY IN GENOMICS DATA
No full textThe ability to sequence the genomic information that describes individual cell states has provided enormous insight into biological systems. However, to sequence the genomic information within a cell, the cell must be killed, preventing measurements from the future states that cell would have occupied had it been allowed to survive. Thus, sequencing measurements only provide a single snapshot in time of cellular genomic states. Often the ultimate goal of an analysis is to derive mechanistic insight into the biology of a system or process from the data. However, such mechanistic, causal inference is almost impossible without temporal information because causality in standard formulations is based on the concept of connected causes and effects through time.
This thesis has interacted with time in genomics data in several ways. The first contribution of this thesis is a neural network-based model that attempts to predict future single-cell transcriptomic states from single-cell transcriptomics data sets. This work demonstrates that using metabolic labeling data sets, future RNA states are estimable within the same cell in the short term, providing a proof of principle that can be expanded as genomics data sets with a temporal dimension become more common.
The second contribution of this thesis is a simulation of molecular cell states over time, which is able to demonstrate how single time points from cells do not allow for robust mechanistic inference. Further, the simulation conforms to observations that mRNA expression and expression of the corresponding protein are often poorly correlated and provides mechanistic explanations for how this occurs.
The final contribution relates to time in a different sense, analyzing the impact of human age on biomarkers used for cancer immunotherapy. We found that older individuals possessed a number of favorable biomarkers at higher levels than their younger counterparts, possibly explaining clinical observations that older individuals do no worse than younger individuals on immune checkpoint therapies despite the usual anticorrelation between patient age and effective immune responses
ESSAYS IN INTERNATIONAL FINANCE AND MACROECONOMICS
No full textThis dissertation contains three essays on global capital flows and public debt. The first chapter studies the role of global banks in cross-border gross and net capital flows. I propose a tractable multi-country model in which leverage-constrained global banks intermediate funds between local banks with heterogeneous projects. Following a relaxation of their constraint, global banks reallocate more funds, generating higher gross capital flows. I show, both theoretically and empirically, that countries with higher net external liabilities to global banks experience a larger deterioration in their current account balance, driven by a larger increase in investment, after a leveraging up by global banks. As such, fluctuations in global banks' leverage also play a key role in driving global imbalances.
The second chapter studies the history of the U.S. public debt/GDP ratio since 1946. We examine the roles of primary budget surpluses, surprise inflation, and pegged interest rates before the Fed-Treasury Accord of 1951 in driving this ratio. Our central result is a simulation of the path that the debt/GDP ratio would have followed with primary budget balance and without the distortions in real interest rates caused by surprise inflation and the pre-Accord peg. Our findings imply that, over the last 76 years, only a small amount of debt reduction has been achieved through growth rates that exceed undistorted interest rates.
The third chapter reexamines the case for growth-indexed bonds (GIBs) by quantifying their impact on the likelihood of reaching high public debt ratios. Although this impact varies across countries and indexation schemes, empirical estimates show a limited reduction in the upper tail of the distribution under the realistic assumption that 20 percent of the stock of debt is indexed to growth. Moreover, a sustained premium of 100 basis points would actually increase the upper tail of the distribution for most countries. Thus, the debt stabilization benefits of GIBs seem elusive unless they are issued at both a substantially large scale and low premium
PAST MEDICAL HISTORY
No full textA physician who writes about medicine has a conundrum to solve: who are you when you write? We may see ourselves as scribes, as translators of facts, or as observers with inside knowledge. But perhaps we should examine our own part in the story.
When we write the patient record, by tradition we include a list: the past medical history. It is both our knowledge of body, and our body of knowledge. This thesis is my own lore, my history, and my knowledge of science and the medical world I inhabit
DEVELOPMENT OF A MACHINE LEARNING MODEL FOR LIVER TRANSPLANTATION
No full textThere are far fewer donor livers available for transplant than patients who need them. This has led to the use of marginal livers – livers that are riskier than typical donor livers yet might still provide a benefit to carefully selected patients. However, the decision to use a particular marginal liver for a particular patient is based largely on clinical gestalt or traditional clinical studies often using regression analysis, which likely do not fully account for the complex relationships between donor, recipient, and transplant characteristics. With the continued advancement of technology and supervised learning computer algorithms, machine learning (ML) methods have emerged as a valuable means of leveraging rich database information to generate personalized predictions (Chapter 1). This dissertation focuses on the utilization of ML algorithms to optimally inform decision making around marginal organ transplantation and enhance patient-physician communication.
We began by leveraging Scientific Registry of Transplant Recipients (SRTR) national data and employing several machine learning techniques to determine which patient characteristics held the most importance in predicting their survival on the waitlist. We then took this output and constructed a waitlist survival model (Chapter 2). Using the same database and machine learning methods, we developed a model that would predict post-transplant survival for a specific patient-liver pairing (Chapter 3).
We then interviewed liver transplant candidates, recipients, and healthcare providers to ascertain stakeholder priorities in designing a decision aid that displays the two aforementioned survival predictions (Chapter 4). Lastly, as all these efforts would be in vain if we could not improve upon long-term survival of the organ, we sought to gain insight into the post-transplantation patient experience and barriers to immunosuppression medication adherence (Chapter 5). We conclude with a summary of significant findings and plans for future research (Chapter 6)
The Study of the SCF E3 Ubiquitin Ligase in the Ceanorhabditis elegans Germline
No full textFaithful chromosome segregation during meiosis requires chromosomes to pair and recombine with their homologous partners during meiotic prophase I. In most eukaryotes, homologous chromosome alignment is reinforced by synapsis, a process defined by the assembly of the synaptonemal complex (SC), a tripartite protein structure that assembles between homologous chromosomes. These processes of pairing, synapsis, and meiotic recombination are tightly regulated and aligned with different phases of the cell cycle. One defining aspect of the cell cycle is its irreversible nature, which is made possible only by the protein degradation that occurs as the cell progresses from one phase to the next. Understanding how protein degradation controls pairing, synapsis, and meiotic recombination allows us to understand how protein degradation regulates the meiotic cell cycle. To investigate the ways in which protein degradation controls meiotic processes, I studied the nematode Caenorhabditis elegans and the E3 ubiquitin ligase, the SCF (SKP1-Cullin-F-box) complex. I used a mix of genetics, cellular biology, and biochemistry to identify new interacting proteins, possible targets, and potential regulatory subunits of the SCF within C. elegans. I found that SCFPROM-1 targets the protein phosphatase PPM-1.D at the onset of meiotic entry, which in turn releases the kinase CHK-2, the master regulator of early meiotic processes. I also found that two paralogous SKP1 proteins moonlight as necessary structural components of the synaptonemal complex. These proteins have evolved to utilize the binding interfaces between SKP1-Cullins and SKP1-F-box proteins to interact with themselves to form a dimer or to interact with the surrounding SC proteins. In addition to identifying these proteins as SC components, we have shown that they are the last two necessary SC components, completing the essential set and allowing for in vitro reconstitution of the SC. Overall, this work provides a better understanding of how protein degradation and its machinery regulate the meiotic cell cycle and its various processes
A BIOETHICS APPROACH TO EXAMINING INADEQUATE ACUTE PAIN CONTROL DURING INTRAUTERINE DEVICE INSERTION
Get PDFIntrauterine device (IUD) insertion procedures can cause pain, where some patients even have an extreme level of pain. This is a health care problem that should prompt further questioning and ethical analysis. While ethics-related concepts have been described in the literature surrounding IUD insertion procedures, there is a shortage of explicit ethical analysis through the use of ethical principles. The purpose of this thesis is to illustrate how bioethics concepts, such as beneficence, nonmaleficence, respect for autonomy, and justice, can serve as a useful lens for examining issues related to pain with IUD insertion procedures. In this analysis, concerns with IUD insertion procedures are grouped and analyzed through the perspective of each ethical principle. Additionally, it is possible to examine how ethical principles conflict with each other within the space of IUD insertions. By assembling a comprehensive review of IUD insertion pain issues through the language of ethics, this thesis brings this reproductive health problem further into the academic bioethics sphere, highlights how the current handling of IUD insertions is ethically problematic, and argues why bioethics should dedicate space and consideration to this topic
Regulation of cytosolic proteostasis and misfolded protein import into mitochondria
No full textMitochondria are the cellular energy hub and central target of metabolic regulation. They also facilitate protein homeostasis (proteostasis) through pathways such as the ‘mitochondria as guardian in cytosol’ (MAGIC) in which cytosolic misfolded proteins are imported into and degraded inside mitochondria. However, the regulation of MAGIC is not well understood, and how the metabolic and proteostatic function of mitochondria is properly balanced remains unclear.
We conducted an unbiased genome-wide screen to identify potential regulators of MAGIC in budding yeast. Snf1, the yeast AMP-activated protein kinase (AMPK), inhibits the import of cytosolic misfolded proteins into mitochondria while promoting mitochondrial biogenesis under glucose starvation. This inhibition requires a downstream transcription factor regulating mitochondrial gene expression and is likely to be conferred through substrate competition and mitochondrial import channel selectivity. Snf1/AMPK activation also protects mitochondrial fitness in yeast under stress induced by misfolded proteins associated with neurodegenerative diseases.
Our genetic screening also uncovered other potential regulators of MAGIC in yeast. Loss of Gas1, a β-1,3-glucanosyltransferase required for cell wall integrity, inhibits the accumulation and degradation of misfolded proteins in mitochondria, while Snf1 remains inactive. By contrast, Gas1 deficiency elevates polyubiquitination and promotes proteasome-mediated degradation. Interestingly, the carboxy-terminal glycosylphosphatidylinositol (GPI) anchor signal of Gas1 localizes to mitochondria after being cleaved in the endoplasmic reticulum (ER), but this mitochondria-associated GPI-anchor signal peptide is not required for MAGIC. Taken together, this study provides a paradigm for studying genes and pathways that affect cytosolic proteostasis upon different environmental or genetic perturbations
INVESTIGATIONS OF MAGNESIUM OXIDE UNDER EXTREME PRESSURE AND TEMPERATURE CONDITIONS
No full textMagnesium oxide (MgO) is a major component of the mantles of Earth and terrestrial planets, playing an important role in understanding their structures and compositions. While MgO has been extensively studied, its properties under extreme high-pressure and high-temperature conditions remain to be further investigated. Recent attention focuses on its phase transition from the B1 (NaCl-type) structure to the B2 (CsCl-type) structure and its melting characteristics. The key to studying the phase transition and equation-of-state of MgO is reaching the high-pressure and high-temperature conditions that encompass expected conditions for the B1-B2 and melting transitions (e.g., 300-800 GPa, 3000-15000 K) and having high quality diagnostics. Conventional static compression techniques, e.g., Diamond Anvil Cells, are usually limited to pressure below 200 GPa and temperatures below 3000 K. Only dynamic compression can drive materials to terapascal pressures and to temperatures beyond 10,000 K by intrinsically increasing internal energy of the materials with shock waves. High-energy lasers have demonstrated significant success in dynamic compression experiments by delivering precisely controlled, high-energy pulses, thereby enabling the generation of ultrafast, well-regulated shock waves, and offering new perspectives on material behavior under extreme conditions.
This thesis aims to elucidate the properties of MgO under high-pressure and high-temperature conditions, including the shock equations of state, the B1-B2 and melting transitions, optical properties, and crystallographic structures under extreme conditions. Specifically, the crystal orientation effects on the properties mentioned above within the experimental timescales will be investigated for the first time.
The thesis is organized into eight chapters, the first two of which provide background information and descriptions of shock compression and diagnostics employed in the shock experiments in this study. The third chapter presents an experimental study on temperature measurements of single-crystal MgO [100] using decaying shock compression. The fourth chapter presents the comprehensive shock compression experiments on single-crystal MgO [110], combining both pyrometry and x-ray diffraction which allow us to peek at the kinetics effect on phase transformations. Chapter five exhibits comprehensive experiments on single-crystal MgO [111] and demonstrates the significant orientation dependency of the high-pressure, high-temperature properties for MgO. The sixth chapter investigates the uncertainties in shock compression experiments to describe the measurements more accurately in shock experiments. Chapter seven and eight are conclusions and appendices, respectively.
This comprehensive investigation of MgO under extreme P-T conditions will expand our knowledge of its properties and reveal orientation-dependent transition kinetics and transition pathways, providing constraints to accurate equation-of-state data for modeling planetary interiors
Selection and Evaluation of Non-invasive Real-time Metrics for Prediction of Graft Survival in Vascularized Composite Allotransplantation
No full textA critical opportunity in the advancement of machine perfusion as an organ and tissue preservation strategy is its potential as a monitoring tool in addition to its role as a source of support for transplants ex vivo. The ability to detect biomarkers present in perfused fluids has the potential to allow researchers to draw inferences about a graft’s health and its likelihood of surviving transplantation. In the field of vascularized composite allotransplantation (VCA), however, it is rare for pre-clinical studies evaluating preservation strategies in small animal models to include the transplantation of grafts; this makes it impossible to determine by perfusion metrics alone whether a graft will survive transplantation. Additionally, many of the biomarkers used in machine perfusion research for VCA are evaluated by invasive means or through assays that cannot provide real-time output, both of which are not conducive for providing feedback regarding graft health.
In this report, the selection of five metrics which are collected non-invasively and in real-time is described. These parameters were evaluated in the context of rat abdominal wall perfusion followed by transplantation. Additionally, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) is explored in this report as a surrogate biomarker for graft viability in the place of transplantation. A major result of our study is the survival of rat abdominal wall grafts following machine perfusion for up to 12 hours, exceeding the expectation of how long a graft can survive in an unoxygenated, warm ischemic environment. Through this study, the selected NIRT metrics were tracked over the course of perfusion in order to correlate variations in parameter values over time with the transplantation outcome. Our results show the promise of our rat abdominal wall perfusion protocol for future VCA transplantation research and clinical applications
CREATION AND ANALYSIS OF A PROCESS MODEL FOR THE DELIVERY OF ANTIBIOTICS TO CHILDREN WITH PRESUMED SEPSIS
No full textSepsis is a life-threatening medical emergency where the timely administration of antibiotics within a one-hour time constraint is crucial for patients' survival. However, this requirement is challenging to meet at the operational level. The purpose of this study is to identify and address potential delays in the order-to-administration time of antibiotics at Johns Hopkins Hospital and propose measures to enhance operational efficiency and shorten the antibiotic delivery delay.
A mixed-methods research design was used, integrating both qualitative and quantitative methods. Clinical interviews and observations were conducted to gain qualitative insights into the process and identify potential bottlenecks. A granular BPMN-style process model was then developed, outlining all the necessary steps in the process at the operational level. Quantitative data analysis of electronic medical records (EMR) revealed that more than half of Cefepime orders were not delivered within the one-hour window.
To enable a retrospective review of the process, we demonstrated the feasibility of constructing a care process for presumed septic patients using event logs and process mining tools. Finally, potential bottlenecks were identified, and corresponding mitigating solutions were proposed. We expect the proposed measures can potentially enhance operational efficiency and reduce antibiotic delivery delay in the order-to-administration time.
This study contributes to the field of healthcare operations, by providing a detailed process model and a method to retrospectively generate the data-oriented process map of actual process flow, to enhance the efficiency of antibiotic delivery at operational level. Our work demonstrates the value of integrating qualitative and quantitative methods to gain a comprehensive understanding of healthcare processes and identify potential improvement opportunities. Ultimately, our findings have potential implications for healthcare providers, system designers for the development of effective interventions to enhance the quality of care and a more efficient antibiotic delivery process