UC Santa Cruz

Ribosomes are essential macromolecules that convert mRNA to polypeptides and functional proteins. Ribosomes can be regulated by multiple mechanisms, including limiting the biogenesis of ribosomal subunits, sequestering 70S ribosomes, dimerizing 70S complexes into translationally inactive 100S ribosome, and “silencing” the 50S subunit to prevent the assembly of active 70S ribosomes. In this review, we focus on the protein that carries out ribosomal silencing, called ribosomal silencing factor S (RsfS), that has been shown in multiple species to bind to the L14 ribosomal protein in the 50S subunit and prevent 50S-to-30S subunit joining. RsfS is a highly conserved small protein found across most bacterial species and in eukaryotes. The direct target of RsfS, L14 (RplN), is one of the most conserved ribosomal proteins. RsfS has been primarily studied for binding interactions with L14 and the downstream effects of preventing 70S ribosome assembly. rsfS null mutants exhibit defects in in vitro (stationary phase, biofilm, growth, nutrient limitation, oxidative stress) and in vivo (macrophage and murine models) conditions in several bacterial species. However, other topics remain underexplored, highlighting areas for future study such as regulation of RsfS and L14 expression, what factors dictate ribosomal interactions, roles in antibiotic tolerance, and whether RsfS has additional physiological roles.

Working between the fields of visual studies and environmental humanities, this dissertation examines works of contemporary visual art and culture that probe the relationship between ecological crises and the production of global and planetary memory. Through analyses that span social practice art, installation, performance, film, museum display, architecture, and sculpture among others, each chapter—on seeds, ice, and whales, respectively—considers the varied social, political, and environmental implications of global and planetary memory practices “at the end of the world.” Geographically centered primarily—though not exclusively—in the Arctic, the dissertation begins by investigating the role of spatial imaginaries and colonial legacies in shaping institutionalized global memory practices of conservation, preservation, and archiving while analyzing artists’ engagements with sustained and reclaimed forms of cultural memory transmission. The project then conceives of various life forms as bearers of memory involving centuries—if not millennia—of environmental and cultural information. In doing so, I argue that “planetary archives” offer novel temporal and spatial ways of producing historical knowledge. Methodologically, my dissertation contributes to visual studies scholarship—particularly scholarship on contemporary art and climate change—by positing nonhuman modes of perception as a means of advancing formal visual analysis—and by advocating for creative approaches to scholarly writing through its use of essayistic modes and the epistolary form. Ultimately, my dissertation contends that through the investigation of contemporary visual-cultural works engaging with emergent memory practices (whether through representation, enactment and/or critique), this era of climate catastrophe warrants an investigation of the expanded role of the “planetary” and the political possibilities afforded to us when we remember differently.

The chemistry of the heavier pnictogens has shaped medicine for centuries, from the earliest therapeutic applications to the modern use of pnictogen compounds in treating infectious disease and the emergence of radioisotopes for diagnostic imaging and targeted radiotherapy. Despite this rich history, many aspects of heavier pnictogen chemistry remain incompletely understood, particularly the relationships between molecular structure, electronic properties, and chemical reactivity in complex biological environments. Addressing these fundamental questions is essential for the rational development of next-generation pnictogen-based therapeutics, diagnostics, and ligand design.This dissertation investigates fundamental aspects of arsenic, antimony, and bismuth chemistry through four complementary studies spanning historical medicinal chemistry, redox reactivity, coordination chemistry, and ligand design. First, the long-debated structure of Ehrlich's arsphenamine is revisited through a critical examination of historical evidence and modern structural characterization, providing a new perspective on one of medicinal chemistry's most influential compounds. Second, a series of arylarsonic acids are examined as redox-activated prodrugs to elucidate how electronic structure governs reduction kinetics, ligand exchange, and biological activity, establishing structure-reactivity relationships that inform the design of modern arsenic therapeutics. Third, the synthesis and characterization of small molecule tricysteinylpnictines are explored as simplified models for understanding the coordination environments of arsenic, antimony, and bismuth in sulfur-rich proteins, providing insight into the structural principles that underlie biological recognition and reactivity. Finally, principles of hard–soft acid– base chemistry are applied to the development of new chelators for heavy pnictogens, with the goal of enabling the stable coordination of medically relevant radioisotopes for future imaging and therapeutic applications.Collectively, this dissertation illustrates that common themes link seemingly disparate problems in heavier pnictogen chemistry. These studies establish fundamental structure-reactivity relationships that provide a foundation for future advances in medicinal, biological, and radiochemical applications of the heavier pnictogens.

A ubiquitous task in statistical practice is that of variable selection, identifying which of a large set of features are the relevant ones. As data sets with a large number of observations have become increasingly common, new theoretical and computational challenges for model selection have emerged. We consider the variable selection problem for linear models in the ℳ-open setting, where the data generating process is outside the model space. We focus on the novel problem of model superinduction, which refers to the tendency of model selection procedures to select larger models at an exponential rate as the sample size grows, resulting in overparametrized models which collapse model interpretability and induce severe computational difficulties. For a set of popular frequentist information criteria and the Bayesian case of mixtures of 𝑔-priors, we prove that when comparing nested models, the larger model will always be asymptotically selected. We seek to minimize this effect for large 𝑛 while preserving variable selection consistency. We propose utilizing a mixture of 𝑔-priors, where the hyper-prior on 𝑔 has hyper-parameters chosen to result in a slowly diminishing rate of prior influence on the posterior, which favors simpler models while preserving consistency. We also propose a model space prior which induces stronger model complexity penalization for large sample sizes. The posterior model probabilities under our prior choices further provide an alternative information criterion that is resistant to the effects of model superinduction. Next, we extend our results to other classes of popular variable selection priors, the family of spike and slab priors, the non-local priors, and selection procedures that correspond to posterior modes such as the LASSO. We show that these procedures are all afflicted with model superinduction, except for the continuous spike and slab priors when a Student-t distribution is used for both the spike and the slab. Finally, we address existing bottlenecks in the computation efficacy of spike and slab based variable selection. We demonstrate that Gibbs samplers scale poorly to large sample sizes, and the proposed alternatives in the literature result in posterior surrogates that are afflicted with superinduction. Instead, we propose a search strategy based off easy to compute approximations of the posterior model probabilities. We show this procedure, Fast Approximate Stochastic Search (FASS), coupled with post-hoc inference on parameters via either a Block-Variational Bayes approach or an Expectation Propagation approach, results in competitive performance. We demonstrate the aforementioned phenomena, and the efficacy of our proposed solutions, via synthetic data examples and case studies using albedo data from GOES satellites and LCA application data.

Human disturbance is one of the most pervasive forces shaping wildlife behavior, and there is growing recognition that human presence itself acts as a distinct form of disturbance. Outdoor recreation is an increasingly popular form of human presence with potential impacts on wildlife, particularly large carnivores whose large spatial and energetic requirements frequently bring them into close contact – and conflict – with people. As interest in outdoor recreation grows, so too does the need to understand it’s impact on wildlife behavior so that we can develop management strategies that support the interests of both people and wildlife. In my dissertation I explore how outdoor recreation impacts mountain lion (Puma concolor) behavior and how we can apply our understanding of human-mountain interactions to develop effective management strategies that promote human-carnivore coexistence. First, I quantify mountain lion response to recreational activity using a novel metric of recreation intensity. I then explore individual variation in mountain lion response to recreation and ask if differences in individual behavioral responses scale up to affect the risk of human-mountain lion conflict (Chapter 1). Next, I investigate the efficacy of two novel behavioral modification strategies that aim to reduce the potential for human-mountain lion conflict on recreational landscapes (Chapter 2). These tools offer non-lethal alternatives to carnivore management in high-human use areas and attempt to reduce potential conflict by reducing spatiotemporal overlap between people and carnivores. Finally, I apply our findings to develop a human-mountain lion interaction management plan that synthesizes what we know about mountain lion ecology to develop management recommendations that support the long-term population persistence of mountain lions in the Santa Cruz Mountains (Chapter 3). This work attempts to integrate novel research findings with applied management practices to support the long-term conservation of carnivore populations living on human-dominated landscapes.

Planetary exploration requires extracting physically meaningful information from incomplete measurements acquired under environmental, instrumental, and operational constraints. This dissertation integrates Raman spectroscopy, X-ray diffraction (XRD), mineral physics, and computational analysis to expand the information recoverable from measurements of Mars-relevant materials. Raman spectroscopy provides the central connection among these studies because its vibrational frequencies, intensities, and spectral shapes encode not only mineral identity, but also bonding, hydration, crystallinity, temperature, structural disorder, and crystallographic orientation. Three studies progressively examine mineral phase identification, mineral physical state, and how it is crystallographically organized. A convolutional neural network trained using high-volume, wide within-class Raman datasets and demonstrated robust mineral classification under dusty, low-signal, and otherwise nonideal conditions relevant to Mars. These results establish that physically representative training populations and expert-guided dataset construction are critical to reliable machine-learning interpretation. Chapter 2 included variable-temperature Raman spectroscopy and XRD characterization of progressive dehydration of magnesium perchlorate and magnesium sulfate from -70 to 400°C. Mg(ClO₄)₂ undergoes discrete crystalline hydration transitions, whereas MgSO₄ progressively develops poorly crystalline and X-ray-amorphous products while retaining Raman sensitivity to local sulfate environments. Wilson GF force constants quantitatively connect vibrational frequencies to hydration-dependent changes in tetrahedral bond stiffness and coupling. A previously unreported Fermi resonance in crystalline Mg(ClO₄)₂ provides an additional hydration diagnostic and potential Raman thermometer. These results demonstrate that full-spectrum interpretation and quantitative vibrational analysis can constrain hydration, bonding, crystallinity, metastability, and environmental history beyond mineral identification alone. Chapter 3 investigates co-registered Raman and synchrotron XRD rasters of Martian shergottite NWA 13276 using combined semi-supervised, human-in-the-loop mineral identification with E-WIMV texture reconstruction. Olivine exhibits strong non-random lattice preferred orientation, while its ~810–870 cm⁻¹ Raman doublet varies systematically with crystallographic orientation, establishing Raman intensity as an orientation-sensitive observable with potential applications to Perseverance rover SuperCam measurements on Mars. This dissertation advances planetary spectroscopy from mineral classification toward physically constrained inference, demonstrating how independent crystallographic measurements, systematic experiments, mineral physics, and expert-guided computation can transform spectral variability from an analytical complication into information about mineral state, structure, and geological history. La exploración planetaria requiere extraer información físicamente significativa de mediciones incompletas adquiridas bajo restricciones ambientales, instrumentales y operativas. Esta tesis integra espectroscopia Raman, difracción de rayos X (XRD), física de minerales, y análisis computacional para ampliar la información recuperable de mediciones de materiales relevantes para Marte. La espectroscopia Raman proporciona la conexión central entre estos estudios porque sus frecuencias vibratorias, intensidades, y formas espectrales codifican no solo la identidad mineral, sino también los enlaces, la hidratación, la cristalinidad, la temperatura, el desorden estructural, y la orientación cristalográfica. Tres estudios examinan progresivamente la identificación de fases minerales, el estado físico de los minerales, y cómo se organiza cristalográficamente. Una red neuronal convolucional entrenada utilizando conjuntos de datos Raman de gran volumen, con amplia variabilidad intraclase y demostró una clasificación robusta de minerales bajo condiciones polvorientas, de baja señal, y otras condiciones no ideales relevantes para Marte. Estos resultados establecen que las poblaciones de entrenamiento físicamente representativas y la construcción de conjuntos de datos guiada por expertos son fundamentales para una interpretación fiable mediante aprendizaje automático. El Capítulo 2 incluyó espectroscopia Raman a temperatura variable y caracterización mediante XRD de la deshidratación progresiva del perclorato de magnesio y el sulfato de magnesio desde -70 hasta 400°C. Mg(ClO₄)₂ experimenta transiciones discretas de hidratación cristalina, mientras que MgSO₄ desarrolla progresivamente productos poco cristalinos y amorfos a los rayos X mientras conserva la sensibilidad Raman a los entornos locales del sulfato. Las constantes de fuerza Wilson GF conectan cuantitativamente las frecuencias vibratorias con cambios dependientes de la hidratación en la rigidez y el acoplamiento de los enlaces tetraédricos. Una resonancia de Fermi no reportada previamente en Mg(ClO₄)₂ cristalino proporciona un diagnóstico adicional de hidratación y un posible termómetro Raman. Estos resultados demuestran que la interpretación del espectro completo y el análisis vibracional cuantitativo pueden restringir la hidratación, los enlaces, la cristalinidad, la metaestabilidad, y la historia ambiental más allá de la mera identificación mineral. El Capítulo 3 investiga barridos Raman y XRD de sincrotrón corregistrados de la shergottita marciana NWA 13276 utilizando una identificación mineral combinada semisupervisada, con intervención humana y reconstrucción de textura E-WIMV. El olivino presenta una fuerte orientación preferente de red no aleatoria, mientras que su doblete Raman de ~810–870 cm⁻¹ varía sistemáticamente con la orientación cristalográfica, estableciendo la intensidad Raman como una magnitud observable sensible a la orientación con posibles aplicaciones a las mediciones de SuperCam del rover Perseverance en Marte.Esta tesis hace avanzar la espectroscopia planetaria desde la clasificación mineral hacia la inferencia físicamente restringida, demostrando cómo las mediciones cristalográficas independientes, los experimentos sistemáticos, la física de minerales, y la computación guiada por expertos pueden transformar la variabilidad espectral de una complicación analítica en información sobre el estado mineral, la estructura, y la historia geológica.

In March 2024, a backdoor was discovered in xz Utils, a widely-used open source data compression library present in nearly every major Linux distribution. The attack was discovered days before merging into major distributions, and if this had happened, it would have allowed attackers to execute arbitrary code on millions of systems worldwide via SSH.The success of this backdoor was enabled by two failures. The first was technical: weaknesses in the software supply chain allowed a malicious actor to inject code into a widely trusted release. The second was human: the project's only maintainer, overwhelmed and burned out after years of maintaining critical infrastructure alone, was the target of a multi-year social engineering campaign, in which a malicious actor built trust under a false identity and gradually obtained commit access to the project. This incident shows that software security failures and sustainability failures are not independent: an overburdened, unsupported maintainer is itself an attack surface.Academic and scientific open source software (OSS) faces both of these crises simultaneously. Projects that critical infrastructure depends on are maintained by researchers, students, and faculty who contribute in their spare time, without dedicated security training or institutional support. Existing security frameworks such as NIST's SSDF, OWASP's SCVS, and SLSA were not designed with these communities in mind, and policy efforts such as the EU Cyber Resilience Act have shown that mandates developed without community input risk harming the ecosystems they are meant to protect.This dissertation addresses the sustainability and security of academic open source software through two parallel empirical research tracks. The sustainability track combines GitHub's REST API with LLM-based filtering to discover and characterize over 216,000 institutionally affiliated repositories across 32 academic and research institutions, finding that while 84% include a README, only 23.4% carry a detectable license and fewer than 2% include a Contributing Guide. Building on this dataset, we develop a maturity-staged sustainability framework that classifies projects into four lifecycle stages and generates targeted recommendations for Open Source Program Offices (OSPOs).The security track examines whether post-9/11 trade security programs offer a workable model for OSS supply-chain policy, finding that effective frameworks require voluntary incentives and direct community engagement rather than top-down mandates. We further evaluate five large language models on the OWASP Benchmark for vulnerability triage, finding that o1-mini reduces false positives by 20% over the Semgrep baseline, demonstrating the potential for automation to reduce the security burden on individual maintainers.Together, these contributions treat sustainability and security as interconnected problems. A project that cannot sustain itself cannot secure its code, and this dissertation takes steps toward closing both gaps.

This thesis forms part of a collaborative project to establish the existence of nontrivial smooth extrinsic biharmonic maps from ℝ4 into 𝕊𝑛 through a min–max construction based on biharmonic replacement. After applying stereographic projection and removable singularity, such maps correspond to extrinsic quasi-biharmonic maps from 𝕊4 into 𝕊𝑛 . This approach is particularly relevant when 𝑛 ≥ 5, since 𝜋4(𝕊𝑛 ) = 0, and hence every map from 𝕊4 to 𝕊𝑛 is null-homotopic. As a result, classical methods such as the direct method in calculus of variations or heat-flow methods within a nontrivial homotopy class cannot be used to construct nontrivial critical points. This project aims to generalize Colding and Minicozzi’s min-max framework for harmonic maps to the case of biharmonic maps. In this thesis, we establish a compactness result for almost biharmonic maps with bounded energy which forms the essential analytical part of the broader min–max framework. In our setting, almost biharmonicity is formulated through local biharmonic replacements rather than a perturbed Euler–Lagrange equation. Under a uniform energy bound, we prove that the min-max sequence arising from this biharmonic replacement procedure bubble-converges subsequentially to a finite collection of quasi-biharmonic maps from 𝕊4 into 𝕊𝑛.

There are many valid reasons for someone to browse the internet privately, and Virtual Private Networks (VPNs) are one method of achieving this. Although primarily used to provide confidentiality and anonymity by encrypting internet traffic and masking IP addresses, users may choose to use a VPN for multiple purposes. However, this raises questions about whether, and how differently, content across the internet is portrayed to VPN users. VPN providers do not all route their traffic the same way, and this study analyzes these differences by investigating two major VPN providers, NordVPN and Surfshark. We measure and quantify the accessibility of popular websites to users of these two VPNs, comparing outcomes against a non-VPN control connection to determine whether, and to what extent, VPN use affects a user's ability to reach and use ordinary websites. This raises a central question: is VPN-related blocking better explained by the network a VPN happens to use, or by where its exit server is physically located? To investigate this, requests to the Tranco top 10,000 domains were compared between each VPN and a non-VPN control connection to measure blocking at full scale. A second set of experiments isolated exit servers by network (ASN) and by physical location across a smaller domain set, allowing the effects of these factors to be compared. At full scale, NordVPN showed 1.59\% VPN-attributable blocking and Surfshark showed 0.42\%. When exit servers were isolated by ASN and location, block-rate variation did not show a clear difference between the two factors, while the network path taken by a request varied substantially with location. These results suggest that VPN-related blocking is not explained by a single infrastructural factor, but instead depends on multiple properties of the network through which a VPN request is routed, including characteristics of the individual exit, network infrastructure, and geographic location. At full scale, NordVPN showed 1.59\% VPN-attributable blocking and Surfshark showed 0.42\%, measured across the Tranco top 10,000 domains using a single exit per provider. Under infrastructure isolation across five ASNs and five countries, crawling top 200 domains, block-rate variation was similar whether ASN or location was held fixed (2.5\% vs. 4.0\% mean within-group range). Per-provider significance testing further found that only a small number of ASNs showed significant differences in blocking. In contrast, network routing varied substantially with exit location: CDN routing varied by 76\% when location was held fixed, compared with 58\% when ASN was held fixed. These results suggest that VPN-related blocking is not explained by a single infrastructural factor, but instead depends on multiple properties of the specific exit and network path through which a VPN request is routed.

Low-abundance biosensing is constrained not only by the sensitivity of a transducer, but by the complete sequence of molecular transport, capture, recognition, translation, transduction, and statistical inference. This dissertation develops a unified framework connecting conventional surface-based electrochemical biosensors with single-molecule solid-state nanopore sensing. The framework addresses a central mismatch in miniaturized biosensors: nanoscale devices can respond strongly to an individual binding or translocation event while sampling so little area or volume that rare targets at femtomolar-to-attomolar concentrations are unlikely to reach the sensing region within a practical measurement time.To decouple molecular collection from transducer area, target capture is distributed throughout the sample volume using antibody-functionalized magnetic beads. Each captured recognition event is subsequently converted into an inert, approximately 200 nm dielectric surrogate through biomolecule-to-surrogate (B2S) translation. The same surrogate population is then measured through two independent physical readouts. In the digital branch, individual surrogates are counted as they translocate a giant solid-state nanopore. Applied to Zika virus NS1 in serum, this approach reaches an analytical limit of detection of 103 aM while preserving event-level current amplitude, dwell-time, and arrival-rate information. In the ensemble branch, translated surrogates are electrophoretically deposited onto an unfunctionalized screen-printed gold electrode, where they restrict redox-probe access and suppress square-wave-voltammetric current. Applied to SARS-CoV-2 nucleocapsid antigen, this portable electrochemical implementation achieves a 4.56 aM limit of detection, a 16.57 aM limit of quantification, and a 9.8-decade quantifiable range, and demonstrates feasibility with RT-PCR-characterized patient samples.A forward stochastic model links antigen abundance, magnetic-bead capture, transfer efficiency, surrogate generation, deposition, diffusional-shadow blocking, square-wave-voltammetric response, and measurement uncertainty without fitting the reported assay figures of merit. The measured performance falls within the model’s predicted distributions, and sensitivity analysis identifies nonspecific blank-surrogate count and its reproducibility as the dominant remaining constraints. Collectively, the results show that separating capture from transduction and translating rare molecular recognition events into robust physical surrogates provides a common route from ensemble electrochemical sensing to digital single-particle detection.