15064 research outputs found
Sort by
Bacterial Motility Patterns Vary Smoothly with Spatial Confinement and Disorder
No full textIn unconfined environments, bacterial motility patterns directly reflect the internal states of the cell. Bacteria operating a run-and-tumble behavioral program swim forward when in a “run” state, and they are stalled in place when in a reorienting “tumble” state. However, in natural environments, motility dynamics are a convolution of bacterial behavior and physical constraints. Recent investigations showed that swimming through highly confined porous media exhibit extended periods of “trapping” punctuated by forward “hops,” suggesting a potential shift in motility strategy. We introduce a microfluidic device to systematically explore bacterial movement in a range of spatially structured environments, bridging the extremes of unconfined and highly confined conditions. We show that run-and-tumble and hop-and-trap are not distinct locomotive modes, but end points of a continuous spectrum of motility. We present the first unifying framework, “swim-and-stall”, to characterize this continuum of observed motility patterns. We demonstrate that a single control program underlies motility across all environments tested—that is, physical structure alone shapes changes in observed output. Our results establish a quantitative link between behavioral rules and environmental context, and show that can navigate dynamic, complex habitats without reprogramming their motility strategy. This robustness may explain the evolutionary persistence of run-and-tumble behavior in a diverse range of peritrichously flagellated bacteria and inform broader models of active transport in structured media.</p
Human Microbiomes Across Space, Time, and Culture: Case Studies from Contemporary South Asia and Ancient Chile
No full textThis dissertation consists of two projects using microbial genomic data to investigate the relationship between humans and their diets. The first project presents the South Asian MicroBiome ARray (SAMBAR), a population-scale 16S gut microbiome study of 575 adults from ten geographically and socio-culturally diverse South Asian communities. Each community was sampled in ancestral villages and urban centers, enabling controlled comparisons of geography and lifestyle. Relative to global cohorts, SAMBAR microbiomes occupy a distinct compositional space with stronger correlation to geography and community membership than lifestyle. Although urbanization is consistently associated with increased abundance of disease-linked taxa, microbiome responses to lifestyle transitions are largely community-driven, including the acquisition of wheat- and dairying-associated microbial modules in some communities that may facilitate non-genetic adaptation to lactase non-persistence. Microbiome responses to urbanization are heterogeneous even at regional scales, reflecting local culture and geography and underscoring the need for community-specific investigations of health impacts. The second project uses ancient metagenomic methods to investigate Streptococcus mutans, a key cariogenic bacterium that metabolizes dietary carbohydrates into enamel-eroding acid. We analyze ancient oral microbiomes from 51 human tooth samples recovered from the El Olivar archaeological site in Chile, ranging from 1157 to 1538 CE. Taxonomic classification reveals predominantly soil-associated microbial communities, as expected for archaeological remains, but also identify endogenous oral taxa in a subset of samples, including taxa previously associated with ancient oral microbiomes. We authenticate ancient S. mutans DNA in 12 samples, and select eight samples for targeted capture and enrichment of phylogenetically informative and virulence-associated S. mutans genes. Phylogenetic analysis integrating ancient and modern strains reveals limited geographic structure, consistent with the high transmissibility of S. mutans, and places all El Olivar samples in a distinct clade with a geographically diverse sister group. Screening for mutacin genes reveals that all eight El Olivar individuals harbor at least one mutacin, and analysis of an 84-bp dextranase gene fragment places El Olivar sequences in close phylogenetic proximity to other pre-colonial American strains. These results add temporal and spatial resolution to our understanding of S. mutans’s virulence and evolution. Together, these projects highlight the potential of microbial genomics to study human diets and lifestyles on both short and long term timelines
Leveraging Active Reinforcement Learning and Generative Models for Biomolecular Design
No full textBiomolecular design plays a critical role across sectors—enabling advances in drug discovery, materials science, energy, and sustainability—while AI-driven approaches are emerging as a transformative force. Yet realizing reliable AI-driven biomolecular design remains challenging due to the need for controllable, constraint-aware generation, unified multimodal integration, robustness to noisy and sparse data, and efficient optimization under costly experimental feedback and vast design spaces, all while preserving biophysical fidelity. This dissertation unifies these challenges under the CURED framework—Controllability, Unified multimodality, Robustness, Efficiency, and Dependability on biological principles—and addresses them by integrating advanced generative models with active reinforcement learning (RL), grounded in theoretical guarantees and first-principles biology. By tightly coupling generative modeling and reinforcement learning—where generative models serve as oracles or world models to guide exploration and data acquisition, reinforcement learning refines generative policies, and both co-evolve across pretraining, post-training, and inference—this work enables controllable, data-efficient, and biologically grounded biomolecular design. Chapter 2 introduces a bi-hierarchical multimodal protein representation framework that integrates sequence-based protein language models with structure-aware graph neural networks. Through bidirectional hierarchical fusion, it learns biologically informed representations and establishes a strong foundation for downstream generative modeling across protein-level, protein–ligand, and protein–protein interaction tasks. Building on this representational foundation, Chapters 3–6 develop a unified family of GPT-based biomolecular generation frameworks for lead discovery and optimization. Chapter 3 presents DrugImproverGPT, which combines GPT pretraining with structured policy optimization (SPO) post-training to enable targeted molecular property optimization while preserving chemical validity and similarity. Chapter 4 introduces ControllableGPT, a controllable pretraining paradigm that couples a causally masked sequence-to-sequence objective with controllable decoding, enabling precise and interpretable molecular edits for lead optimization. Chapter 5 proposes ScaffoldGPT, a scaffold-centric framework integrating multi-stage pretraining, RL post-training, and decoding-level optimization to achieve robust scaffold-preserving molecular design under biophysical constraints. Chapter 6 further extends this line with FragmentGPT, the first GPT-based model unifying fragment growing, linking, and merging, enabled by chemically and energy-aware pretraining and Reward Ranked Alignment with Expert Exploration (RAE) for diversity and multi-objective optimization. Chapters 7–10 establish a principled foundation for active learning and reinforcement learning with multiple experts. Chapter 7 introduces Contextual Active Model Selection (CAMS) with theoretical guarantees for cost-aware expert querying. Chapters 8 and 9 extend this paradigm to full RL, developing algorithms for active policy selection and a robust self-improvement framework that unifies imitation learning and RL under imperfect oracles. Chapter 10 presents Active Advantage-Aligned Reinforcement Learning (A3RL), bridging offline and online RL through confidence-aware sampling to improve sample efficiency under limited data. Finally, Chapter 11 introduces Entropy-Reinforced Planning (ERP) to enhance inference-time exploration, and Chapter 12 synthesizes prior advances with MCTD-ME, a planning-augmented diffusion framework that unifies masked diffusion, multi-expert learning, RL-based planning, and biophysical principles for scalable protein design. Together, these contributions advance the theoretical and practical foundations of AI-driven biomolecular design and establish a cohesive framework that unifies generative modeling and active reinforcement learning to address the CURED challenges—enabling controllable, multimodal, robust, efficient, and biophysically grounded design. By bridging reinforcement learning, generative AI, and first-principles biophysics, this foundation pave the way for more robust, diverse, and biologically faithful design systems that can accelerate discovery across drug development, protein engineering, and the broader life sciences
Combinatorial Problems Arising from Quantum Computing
No full textThis thesis studies certain combinatorial problems that arise in the study of quantum computation. More precisely, we establish oracle results that exemplify ways in which the behavior of quantum polynomial time () can be remarkably decoupled from that of classical complexity classes like and . We also study a problem related to a conjecture which would imply quantum supremacy results: bounding the cardinality of the range of the permanent
Understanding Neural Variability: Mechanistic and Computational Models of Population Activity
No full textNeural activity in the brain is stochastic in nature, with variable responses across trials under the same experimental conditions. It is also high-dimensional, with thousands of neurons in a network working in concert to fulfill a computational role. This dissertation presents three models that each employ different mathematical techniques to produce theoretical insights on neural variability without reducing the dimensionality of the system as is classically done in mean-field theories. In the first project, we prove in a recurrent circuit model that the more heterogeneous the firing rates of neurons in a population, the lower the effective dimension of their trial-to-trial covariability. This was achieved by using operator-valued free probability theory to analyze the interaction between external inputs and recurrent dynamics. The second project addresses a long-standing question in neuroscience about how the neural code for a sensory, motor, or cognitive variable should be organized to optimize its discriminability. We analytically minimize the average binary classification error of a circular variable in the function space of all population tuning curves for various noise models by solving a nonlocal variational problem. We obtained the solution by viewing the space of neural response distributions as a Riemannian manifold in the sense of information geometry and utilizing a result from knot energy theory. The first two projects both make novel predictions that are verified in experimental data. The third project also concerns discriminability, but addresses the classification of discrete classes instead of continuous variables. Neural manifold capacity is a recently introduced measure of representational geometry that quantifies how many objects a large population can represent while ensuring the feasibility of linear classifications with high probability. We present novel derivations of three versions of the manifold capacity formula based on integral geometry, which introduce a new mathematical perspective to the problem compared to the original derivation based on statistical mechanics techniques.</p
Study of Coherent Synchrotron Radiation Effects Using Generative Phase Space Reconstruction
No full textParticle accelerators are machines of great importance in many scientific disciplines, including physics, chemistry, biology, and materials science. In particular, free-electron lasers (FELs) have revolutionized the study of matter at atomic and molecular scales by providing intense, ultrashort x-ray pulses. From the accelerator physics perspective, the production and transport of high-brightness, ultrashort electron bunches necessary for FEL operation remain central challenges in the development of next-generation facilities. One of the main limiting factors for electron-beam quality in FEL linear accelerators is the coherent synchrotron radiation (CSR) emitted during bunch compression. CSR induces a tail–head self-interaction within the beam via radiation emitted from the tail, which distorts the beam distribution in phase space and consequently degrades its quality. Therefore, understanding the CSR-induced effects on the beam phase-space distribution is essential for the optimal operation of present and future facilities. However, experimental studies of CSR effects generally rely on measurements of one- or two-dimensional projections of the full six-dimensional phase space distribution, which limits the observation of the intricate beam structures produced by CSR. This dissertation presents the first experimental measurement of the six-dimensional phase space distribution of a beam influenced by CSR, conducted at the Argonne Wakefield Accelerator Facility (AWA). To enable this measurement, the generative phase space reconstruction method (GPSR) has been developed, which allows six-dimensional phase space reconstructions with as few as 20 two-dimensional measurements of the transverse beam profile. This work also describes the implementation of differentiable beam dynamics simulations as a core component of the GPSR method. The experimental results suggest the presence of CSR effects for a 1 mm-long, 1 nC beam at the AWA reverse chicane section, and the methodological advancements presented here lay the foundation for experimental studies of CSR effects using GPSR. </p
Harnessing Genetic Tools and Metabolic Pathways of Gut Microbes to Enhance Host Resilience Against Pathogens, Inflammation, and Metabolic Disease
No full textThe intestinal microbiota profoundly shapes host physiology, influencing metabolic, immune, and barrier functions. However, the mechanistic dissection of these effects has been limited by the inaccessibility of many commensal species to genetic manipulation. My dissertation addresses this challenge through the development and application of genetic tools to engineer gut-resident. These efforts enabled the construction of programmable commensals capable of delivering therapeutic effectors and probing microbe–host interactions in vivo. First, this work expands the molecular toolkit for previously intractable anaerobes to achieve stable gene expression across multiple species. Using genetic tools, multiple species were engineered to produce defined metabolites and cytokines, including recombinant IL-22 and tryptophan-derived aryl hydrocarbon receptor (AhR) ligands, thereby enabling targeted modulation of hepatic and intestinal pathways in diet-induced metabolic disease models. Complementary multi-omics and histological analyses, including metagenomics, transcriptomics, metabolomics, and high-resolution image segmentation, revealed how microbial gene function shapes host metabolic outcomes such as steatosis and inflammatory signaling. Together, these studies bridge microbial engineering, host physiology, and mechanistic microbiome research
The Versified Turkic Translation of al-Wāqidī’s Futūḥ al-Shām by Ibn Ajā
No full textThis contribution presents the versified Turkic translation of the Futūḥ al-Shām or Conquest of Syria of al-Wāqidī composed by Shams al-Dīn Muḥammad ibn Maḥmūd ibn Khalīl al-Ḥalabī al-Ḥanafī, commonly known as Ibn Ajā (820–81/1417 or 1418–76). I begin by sketching the positionality of his mastery of Turkic in relation to his education within Arabic-Islamic scholarly traditions. I also suggest that the two volumes in which this text has been preserved are written in Ibn Ajā’s own hand. Subsequently, I engage with Ibn Ajā’s reworking of this text by discussing paratextual elements, the additions he introduced to the body of the narrative, and the mode of transposition followed in his versified Turkic rendering of the Arabic prose of the original. In this context, I also briefly discuss the problem of comparing a translation of so notoriously an unstable and “open” text as the Futūḥ al-Shām of al-Wāqidī with Arabic renderings of the same topic. In conclusion, I argue that the literary Turkic idiom chosen by Ibn Ajā should not be classified as “Ottoman” notwithstanding its close accordance with the literary Turkic subsequently centered in the Ottoman realms. Instead, I suggest that the attestation of this idiom in a work compiled in the Mamluk realms of the 1470s indicates that this literary language must be described as a form of courtly “pre-sixteenth-century supra-regional southwest literary Turkic” notwithstanding its straightforward intelligibility to modern scholars familiar with its (later) development as “Ottoman” Turkish
The Coptic Grammatical Tradition of the Muqaddimāt
No full textIn this article, I argue that the thirteenth-century grammatical treatises al-muqaddimāt al-khams, although belonging to the same genre and often transmitted together, constitute five individual and innovative texts. The characteristics of each muqaddimah allow us to trace the different stages of the transformation of Coptic into a classical language with its own grammatical tradition. To this end, I analyze the structural features of the muqaddimāt and how they relate to the different stages of development that the Coptic grammatical tradition underwent. The pioneering muqaddimah of al-Samannūdī was followed by the more systematized grammars by Ibn al-ʿAssāl and Ibn Kātib Qayṣar. Ibn Kātib Qayṣar also applied the terminology of the Arabic grammatical tradition more stringently to Coptic grammar. Al-Qalyūbī departs from this method in favor of a more descriptive approach to Coptic and highlights methods of translation from Coptic to Arabic. Ibn al-Duhayrī comments on the grammatical positions of his predecessors in dialectal fashion and thus constitutes the endpoint in the development of the muqaddimah from a brief introduction to Coptic grammar to a scholarly genre