Alliance One Tobacco (Malawi)
Academic Research Repository at the Institute of Developing EconomiesNot a member yet
52305 research outputs found
Sort by
Building Beyond the Standard Model: Tools from Cosmology to Particle Theory
No full textThe current landscape of high-energy theory and cosmology suffers from an abundance of theoretical models coupled with a lack of immediate new data to constrain them. This thesis addresses several open questions in early-universe cosmology and beyond-the-Standard-Model (BSM) particle physics, specifically focusing on inflation, dark matter, axions and axion strings, and baryogenesis. Rather than relying solely on forthcoming experimental results, I argue for refining the BSM model space through theoretical consistency checks and cross-examination against multiple existing datasets. For instance, inflationary models often encounter severe naturalness (or ) problems; applying effective field theory techniques and symmetry considerations helps exclude models that fail these theoretical standards. Similarly, axion models naturally predict distinctive topological defects, such as axion strings, allowing their observational signatures (or lack thereof) to impose valuable constraints. On the observational side, I emphasize that jointly analyzing the Hubble and large-scale structure (LSS) tensions can significantly constrain dark-sector theories. Additionally, I introduce a novel observational probe, analyzing the effect of early universe inhomogeneities generated prior to Big Bang Nucleosynthesis (BBN) on predicted deuterium abundances. This probe rules out baryogenesis scenarios that produce excessive inhomogeneities which are not fully erased by diffusion, and it can potentially constrain regions of parameter space in prominent high energy scale models such as electroweak baryogenesis (EWBG). Together, these theoretical and observational approaches offer robust methods for refining the BSM landscape and demonstrate the critical role particle theorists can play in interpreting cosmological data, even in the absence of new experimental results.Physic
The Politics of Purity: Democratic Transformations in Nepal's Southern Borderland
No full textThis dissertation considers the relationship between endurance and change within political life in Nepal. Over the preceding decades, the country has been part of a regional phenomenon that scholars have called the “democratization of democracy”—a process in which groups historically marginalized in formal democratic politics have mobilized to assert their presence, demand social justice, and claim political power. In Nepal, these struggles were institutionally realized in the 2008 abolition of the monarchy and the establishment of a federal system, developments that many Nepalis heralded at the time as the beginning of “New Nepal,” one no longer based in the inequalities and hierarchies of history. Drawing on data collected from two years of ethnographic fieldwork in the city of Birgunj in the country’s southern plains, the dissertation examines this transformation through an analysis of political life in Nepal over the longue durée. The southern plains region, known locally as the Tarai or Madhesh, has a history of political marginalization, and activists there were at the forefront of struggles for a federal system. The dissertation considers why, in the years after the establishment of a federal republic, there is nostalgia for the king in some circles in Birgunj. The chapters illustrate the ways in which older social and political forms of space, community, legitimacy, and rule have inflected newer political formations and dynamics, shaping how things like democracy, political belonging, accountability, and nationalism are imagined and practiced in contemporary Nepal. In doing so, the dissertation contends that historical orders and imaginaries have a continued resonance in contemporary political life even as they are changed through processes of modernization and democratization.Anthropolog
Task-Relevant Generative Models and Safe Reinforcement Learning with Applications to Clinical Decision-Making
No full textClinical decision-making is inherently complex and high-stakes, frequently requiring sequential decisions under uncertainty.
Automated discovery of clinical concepts and reinforcement learning (RL) have great potential to assist clinicians, yet current approaches face several practical challenges.
First, generative models often fail to differentiate clinically relevant structures from irrelevant noise, compromising interpretability and predictive utility.
Second, traditional RL approaches optimize for fixed objectives and thus cannot adequately accommodate varying clinical goals or patient-specific preferences.
Finally, prevalent offline RL methods tend to be unsafe, overly conservative, or reliant on impractical assumptions, such as access to the behavior policy, which restricts their real-world usability.
This thesis introduces methodologies designed specifically to address these challenges.
First, Chapter 3 introduces prediction-focused Gaussian Mixture Models (pf-GMM) and Hidden Markov Models (pf-HMM) for identifying and clustering clinically relevant features from noisy, high-dimensional data.
Second, Chapter 4 proposes a Robust Decision-Focused (RDF) model-based RL framework. This framework learns transition dynamics that perform consistently well across changing clinical reward preferences, ensuring high-quality decisions in diverse clinical scenarios.
Third, Chapter 5 presents Decision-Point RL (DPRL), an offline RL methodology that identifies high-confidence ``decision points" in clinical data for targeted, minimal policy adjustments, backed by theoretical safety guarantees.
Validation of these contributions includes rigorous theoretical analyses and extensive empirical evaluations using synthetic benchmarks, medical simulators (e.g., cancer treatment), and real-world clinical datasets (e.g., a hypotension cohort from MIMIC-IV dataset, an HIV dataset, and electronic health records from a hospital system).
The methods presented demonstrate improvements in predictive accuracy, clinical interpretability, robustness against reward shifts, and safety compared to conventional approaches.
Collectively, this thesis advances the development of interpretable, robust, and safe machine learning methods, effectively bridging the gap between machine learning methods and real-world clinical practice, enhancing decision-making in healthcare.Engineering and Applied Sciences - Computer Scienc
Regulation of homotypic and heterotypic interactions in transcription factors
No full textFine-tuned regulation of transcription in cells is essential for proper differentiation, development, and signaling. Transcription factors (TFs) represent an important class of transcriptional regulators that recognize specific DNA sequences and contribute to changes in gene expression. TFs engage in various homotypic and heterotypic interactions with protein cofactors that provide opportunities for their regulation. Aberrant TF function is implicated in a wide range of diseases, but small molecule-mediated modulation of TFs remains a challenging task. Recent advances in chemically induced proximity (CIP) and targeted protein degradation (TPD), however, have highlighted novel ways to perturb TF function, stability, and localization. In this dissertation, I use structural, biochemical, and computational approaches to investigate homotypic interactions among the ZBTB family of TFs and drug-induced heterotypic interactions between the E3 ligase CRBN and ZF domain-containing TFs. Our results from the former study indicate that polymerization in ZBTB TFs enhances their transcriptional effector functions and presents opportunities for therapeutic modulation. Moreover, our latter study defines the landscape of ZF domain-containing TFs amenable to CRBN-focused TPD approaches. Taken together, this dissertation augments our understanding of how TF-mediated interactions and their regulation can be utilized for pharmacological benefit.Biological and Biomedical Science
Genomic Surveillance and Deployable Molecular Diagnostics for Emerging Infectious Diseases
No full textEmergence of novel pathogens and outbreaks of existing biothreats have significant social and economic impacts. Detecting emerging infectious disease (EID) threats early and accurately is critical for timely public health intervention and development of vaccines and therapeutics. In this work, we applied unbiased metagenomic sequencing to detect and characterize both viral and bacterial pathogens in plasma samples from a cohort of febrile patients and healthy controls in Thiès, Senegal. We identified relapsing fever Borrelia, an underrecognized tick-borne bacterial pathogen, as the most common cause of non-malarial febrile illness. Second, we took a genomics-informed approach to designing a deployable reverse transcription loop-mediated isothermal amplification (RT-LAMP) assay for Lassa virus (LASV), a seasonal hemorrhagic fever virus endemic to West Africa. We developed a high-throughput system for testing RT-LAMP primer set activity across diverse in vitro transcribed RNA targets. This massive-scale primer set screening generated important insights on the factors affecting RT-LAMP amplification speed and guided our design of candidate RT-LAMP assays for the two most prevalent lineages of LASV, Lineage II (LII) and Lineage IV (LIV). We evaluated the performance of candidate assays on clinical samples and showed they could detect LASV RNA (sensitivity compared to gold-standard qRT-PCR: Broad LII v1.1 45%, Broad LII v2 50%, Broad LIV-Liberia 33%), especially in samples with a high viral RNA load (sensitivity in samples with qRT-PCR Ct 35: Broad LII v1.1 87%, Broad LII v2 77%). Finally, we used our empirical dataset of over 3,800 unique Lassa virus RT-LAMP primer set (LPS)-target pairs to predict amplification in silico with high precision and recall (SS: Precision = 0.887, Recall = 0.873; WS: Precision = 0.952, Recal = 0.714) and explored the potential for expanding our empirical dataset and applying a biological sequence optimized machine learning architecture to create a tool for rapid RT-LAMP assay design for emerging viral threats.Biological and Biomedical Science
Novel drivers of eukaryotic protein biogenesis and complex assembly
No full textThe cell’s capacity for protein folding can be controlled by transcriptional regulation of genes encoding chaperones. These factors recognize unfolded proteins to shield them from aggregation or actively assist their folding process. Protein synthesis is an intrinsic folding burden on the cell that is monitored by the heat shock (transcription) factor 1 (Hsf1) to maintain the appropriate expression of a small number of essential chaperones for general protein folding. The starting point for my first body of work was the lab’s observation that Hsf1 in the yeast S. cerevisiae additionally controls expression of Zpr1, a protein that bears no homology to other chaperones but is essential and conserved across eukaryotes and archaea. Using complementary approaches, including biochemical reconstitution and structure-guided mutagenesis, we found that Zpr1 is a chaperone tailored to the final steps in the biogenesis of eukaryotic translation elongation factor 1A (eEF1A), a highly abundant GTP-binding (G) protein comprising ~5% of the proteome. The extreme fragility of eEF1A’s tertiary structure had been historically appreciated since the 1970s, when eEF1A’s biochemical activity was first characterized. Our work explained how cells efficiently solve this problem to enable rapid growth while staving off the inherent potential of abundant eEF1A folding intermediates to disrupt global protein folding in the cell.
My subsequent work was centered on the identification of two additional factors in yeast, both conserved but uncharacterized, that support eEF1A biogenesis. First, we identified Aim29 by forward genetic screening as a factor that promotes Zpr1’s essential function in the cell. Follow-up work by other lab members showed that Aim29 is a co-chaperone for Zpr1 that facilitates substrate release once eEF1A has acquired the ability to hydrolyze GTP. Second, using an AlphaFold-guided computational screen, we predicted the function of Ypl225w as a chaperone dedicated to folding eEF1A’s N-terminal G domain. Using a myriad of assays, we found that Ypl225w associates with ribosomes in the act of synthesizing short eEF1A nascent chains. Ypl225w then remains stably bound to translating ribosomes until the emergence of the full complement of GTP-binding sequence elements within the G domain of eEF1A. Lastly, GTP binding to nascent eEF1A drives G domain folding while triggering release of Ypl225w, thereby allowing chaperone recycling. Together, this body of my thesis work revealed that cells use a dedicated team of folding factors to guide eEF1A’s biogenesis beginning with its nascency on the ribosome. By contrast to general chaperone ATPases, eEF1A ATP-independent chaperones receive cues about biochemical directionality and folding product quality via GTP binding and hydrolysis of their sole client.
Next, we focused on identification of an assembly factor for the eukaryotic chaperonin TRiC/CCT. Chaperonins are large, ring-shaped complexes that mediate folding of proteins inside their nanocages. The cytoskeletal protein tubulin is an obligate CCT folding substrate that undergoes a series of sequential folding steps while being encapsulated. Each of these steps comprises a stereotypical interaction with one of eight paralogous CCT subunits that exist in a defined spatial arrangement relative to one another. Since individual CCT subunits lack inherent information for correct self-assembly, we used AlphaFold-guided computational screening to identify the missing assembly factor(s). We show that a conserved tubulin-like protein (Dml1 in yeast) functions as a CCT substrate mimic to guide the association of the positively-charged half-ring of subunits, namely Cct6, Cct3, and Cct1. Using an electrostatic hook, Dml1 next links Cct1 to the negatively-charged half-ring of subunits via Cct4, which arrives at this junction associated with subunits Cct2, Cct5, and Cct7. Finally, we show that Dml1 is required for the association of Cct8 to Cct6, marking the closing of the chaperonin rings en route to CCT complex maturation.Biology, Molecular and Cellula
