12023 research outputs found
Sort by
Advancing Stimulated Raman Scattering Microscopy through Deep Learning and Gel-Based Tissue Engineering
Get PDFStimulated Raman scattering (SRS) microscopy is a highly effective label-free imaging method for investigating the molecular composition of biological systems. Its broader use has been held back by spatial resolution, imaging speed, and large-scale tissue imaging compatibility. Breaking these limitations requires an integrated approach beyond the development of optical hardware. This thesis introduces a compilation of techniques that leverage gel-based tissue engineering and deep learning to enhance the capabilities of SRS microscopy.
The first chapter, Gel-Enabled Super-Resolution Label-Free Volumetric Vibrational Imaging, introduces VISTA, a sample-expansion vibrational imaging technique that achieves label-free super-resolution imaging of protein-dense biological structures with resolution as fine as 78 nm. By enabling isotropic expansion and protein retention, VISTA allows for high-throughput, unbiased volumetric imaging without labeling, with further enhancement using deep learning-based component prediction.
The second chapter, High-Resolution Imaging of In Vivo Protein Aggregates, applies VISTA to image amyloid-beta and polyQ aggregates in biological samples with high specificity. Combined with segmentation using convolutional neural networks, this technique is capable of mapping aggregate structure and microenvironments, enabling new insights into neurodegenerative disease pathology.
The third chapter, High-Throughput Volumetric Mapping Facilitated by Active Tissue SHRINK, introduces SHRINK, a hydrogel-based sample shrinkage method that isotropically shrinks tissue while maintaining structural integrity. Active shrinkage enhances imaging throughput and signal sensitivity and enables rapid, large-scale, three-dimensional whole-organ mapping with SRS microscopy.
The fourth chapter, Deep Learning-Augmented Metabolic Profiling in Live Neuronal Cultures, presents a tandem deep learning platform for live-cell metabolic imaging. By integrating a recurrent convolutional neural network and U-Net segmentation model with deuterium-labeled metabolic tracing, this platform enables non-invasive, high-speed profiling of lipid, protein, glucose, and water metabolism in neuronal subtypes under physiological and pathological conditions.
These developments represent multidimensional strategies that expand the application of SRS microscopy to high-resolution, high-throughput, and dynamic imaging in a variety of biological systems. The integration of deep learning and gel-based tissue engineering techniques opens new avenues for SRS microscopy to explore complex biological questions.</p
Essays in Experimental Economics
No full textThis dissertation consists of three essays that use lab and online experiments to investigate how individuals make decisions under uncertainty, in social contexts, and when forming beliefs about others. Each essay introduces a distinct setting, but all share a common goal, which is to improve our understanding of human decision-making.
Chapter 1 examines commitment contracts. Their high rates of failure raise concerns since individuals may end up worse off than if they had never committed. We investigate whether some of these failures are actually anticipated, with individuals recognizing that future uncertainty might make failing the contract the best option upon some realizations of uncertainty. We refer to this behavior as planning for the possibility of failure. This approach is different from the usual interpretation of failures, which we call failing to plan, as it attributes failures to take-up mistakes. To study whether individuals plan for the possibility of failure, we conducted a controlled lab experiment designed to detect patterns of such planning. Our findings indicate that about one-third of all commitment choices can be attributed to this kind of foresight. This suggests that planning for failure is common, and that high failure rates are not necessarily driven by mistaken commitments. Thus, they do not by themselves call into question the value of commitment contracts.
The second essay studies the decision to ask for help—a behavior that can be critical in addressing information asymmetries but is often avoided. In an online experiment, we find that making potential helpers even minimally identifiable (e.g., through an uninformative ID number) significantly increases the likelihood of asking. Belief data suggest that this effect stems from shifts in how individuals weigh expected payoffs and other factors (particularly social ones) when deciding whether to ask.
The third essay explores how people expect others to update their beliefs upon receiving new information. We find that when two individuals have different priors, people expect others’ beliefs to move toward their own prior upon receiving new information. Although this result is consistent with the theoretical predictions for Bayesian agents, we find no support for the precision of information affecting the magnitude of the shift in the way the theory predicts. We find that this effect occurs not only due to under-updating of one's own beliefs but also due to recognition of under-updating by others.</p
Structural and Mechanistic Studies of Membrane Protein Biogenesis and Quality Control at the Endoplasmic Reticulum
No full textMembrane proteins make up around 30% of the human proteome and carry out essential functions in the cell, including but not limited to signaling, cell adhesion, and metabolic transport. To maintain homeostasis, the eukaryotic cell has evolved complex pathways for membrane protein biogenesis and quality control, as they contain hydrophobic transmembrane domains (TMDs) that can easily aggregate in the aqueous environment of the cytosol causing cellular damage. At the endoplasmic reticulum (ER), the ER membrane protein complex (EMC) co-translationally inserts the first TMD of multipass membrane protein and post-translationally inserts tail-anchored membrane proteins. In this thesis, we are able to show how the EMC can coordinate with the other protein machineries of the multipass translocon (such as the back-of-Sec61/BOS complex) at the ER to accommodate the insertion of diverse multipass membrane protein substrates, depending on the biophysical properties of their N-terminal soluble domain. We also structurally characterize the EMC•BOS holocomplex, highlighting its spatial relation with the other biogenesis factors at the multipass translocon. In addition, this thesis also explores how a novel quality control factor TXNDC15 functions in ER-associated degradation (ERAD) to facilitate degradation of unassembled membrane proteins. We found that TXNDC15 works with the E3 ligase MARCHF6 and recognizes a set of membrane proteins with exposed hydrophobic domain in the ER lumen. Lastly, we were able to also link TXNDC15’s quality control function to its implication in ciliopathies, such as Meckel-Gruber syndrome
Low-Overhead Quantum Fault Tolerance
No full textFault tolerance is an essential property of future quantum computers where a quantum computation is mapped to a new one that is resilient to operational errors. This resilience comes at an additional time and space overhead. In this thesis, we study schemes that asymptotically reduce the overhead of quantum fault tolerance in various models of computation. We construct a scheme for fault-tolerant quantum computation that requires nearly-logarithmic spacetime overhead assuming access to noiseless classical computation. In the second half, we construct a quantum memory with a memory threshold that uses local quantum operations on a 2D lattice
Resolvent Analysis of Non-Stationary Turbulent Flows and Transient Flow Phenomena
Get PDFIn this work, we develop a wavelet-based formulation of resolvent analysis in order to extend the method to transient phenomena and non-stationary flows. We apply this method in two ways: first, to analyze systems that were not previously amenable to traditional resolvent analysis, and second, to probe the limits of the resolvent forcing modes' "optimality" in a nonlinear simulation as well as investigate the mechanisms that suppress their effectiveness. In wavelet-based resolvent analysis, the Navier-Stokes equations are linearized about a mean profile, Fourier-transformed in the homogeneous directions, and wavelet-transformed in time. The nonlinear terms are represented as forcing terms acting on the system, and a maximally perturbing forcing mode and the response it produces are then computed for this linear system. The wavelet formulation enables the forcing and response modes to represent transient trajectories. By windowing the wavelet-based resolvent operator, we can also compute optimal forcing modes restricted to a time-localized pulse along with their transient response.
For the first application of the method, we use the windowing approach to study bursting in channel flow. The optimal response mode grows and decays in time scales that match turbulent data, and we show that this optimal burst exploits the Orr mechanism.
We also study channel flow subjected to a spanwise pressure gradient. The corresponding resolvent modes mirror the mean flow and gradually realign themselves according to the new flow conditions. More interestingly, they exhibit a collapse of the lift-up mechanism during this realignment, which offers an explanation to the depletion of tangential Reynolds stresses in the turbulent system.
For the second application of the method, we inject time-localized resolvent forcing modes for the minimal flow unit into a simulation of the system, at different intensities. The principal resolvent forcing mode is much more effective than a randomly generated forcing structure at amplifying the near-wall streak. For initial times and close to the wall, the turbulent minimal flow unit matches the principal response mode well, but due to nonlinear effects, the response decays prematurely. By computing the nonlinear energy transfer to secondary scales, we find that the breakdown of the actuated mode proceeds similarly across all forcing intensities: in the near-wall region, the induced streak forks into two branches, while in the outer region, the streak breaks up in the streamwise direction. In both regions, spanwise gradients account for the dominant share of nonlinear energy transfer.</p
Investigation and Control of the Electrode/Electrolyte Interface in Electrochemical Systems
Get PDFIn electrochemical reactions, the electrode/electrolyte interface is of vital importance, as no reactivity occurs in the bulk electrode or the electrolyte. Often, the interface can be the difference between a successful reaction and a failure. In this thesis, we present three works wherein the electrode/electrolyte interface is studied and controlled to drive desired electrochemical reactivity. A Mg-In alloy is employed for Mg metal batteries to prevent Mg dendrite growth, which can cause cell shorting and failure. By coating the surface of Mg metal electrodes with the Mg-In alloy, the nucleation of Mg dendrites is suppressed and instead the Mg electroalloys into the surface alloy upon reduction, significantly increasing the cell life time. Next, the Li-intercalation material LiTiS₂ is studied for use in organic reductive electrosynthesis as counter anodes. Traditional metal sacrificial counter anodes are known to cause issues such as surface passivation, chemical reactivity, and cross-plating at the working electrode, which is deleterious to the desired organic reactivity. It is found that LiTiS₂ surface is less reactive in organic electrolytes, reducing both passivation and chemical reactivity. Further, Li⁺ de-intercalated from LiTiS₂ oxidation is found to be less susceptible to cross-plating than Zn, a common sacrificial anode. Finally, the effect of electrode material on the electrochemical reduction of ᵗBuI is studied. Using electrochemical characterization, it is found that the reduction is catalyzed on Au and Ag through adsorption of the initial substrate, as well as the adsorption of the reactive intermediate tBu radical. The catalysis of ᵗBuI reduction can have meaningful consequences for organic reactivity, driving the desirable generation of the carbanion nucleophile from alkyl halide reactants
Studies on Scaling Throughput in Protein Engineering
Get PDFIn this work we present three studies in protein engineering. While all three protein classes that have been targeted for engineering tasks are very different, the studies have a focus on scaling-up the throughput in protein engineering.
The first study concerns machine learning (ML) based antibody humanization techniques. Achieving a reduction of patient anti-drug antibody responses in clinical trials is the goal of antibody humanization. To measure this however, one needs to pass significant scientific, bureaucratic, and financial hurdles, which is very rarely done and especially never at scale. Most existing ML-based antibody humanization techniques claim that they work without providing any experimental evidence. We developed Mousify as an in silico antibody humanization platform to place existing models into one framework for wet-laboratory validation. We demonstrate that even the best models have a fundamental flaw in that they only generate a single antibody. We use Mousify and Markov chains to show that using ML-based antibody humanization models for library generation is not only feasible but produces both stable and functional variants. Learning the lessons from our wet-laboratory experiments, we then developed a variational autoencoder model with properties that hopefully improve the outcomes of antibody humanization experiments.
In the second study, we outline our plans and initial results to develop a bioelectrocatalytic system for the conversion of N2 to ammonia using nitrogenase. Most of the world’s ammonia is used for agricultural purposes and is produced via the environmentally damaging Haber-Bosch process. Engineering nitrogenase for the bioelectrocatalytic production of ammonia is not trivial and a high throughput is not guaranteed. We present preliminary results in how throughput can be increased through diazotrophic pre-selection of nitrogenase variants, as well as a quest to find the ideal starting point for engineering using a combination of ancestral sequence reconstruction and generative protein language models.
In the third and final study we present a directed evolution campaign to evolve protoglobins for the enantioselective catalytic formation of cis-trifluoromethyl substituted cyclopropanes, the first such reaction in both the chemical and biological world. Not only is the enzyme ApePgb LQ capable of efficiently performing carbene insertions into double-bonds, but it also shows a much more diverse substrate scope than similar enantioselective formations of trans-trifluoromethyl substituted cyclopropanes. After demonstrating that ApePgb LQ reactions can be increased to a 1-mmol scale, we investigated the nature of protoglobin cis-selectivity using various computational methods.</p
High-Field Charge Transport and Fluctuation Phenomena in Semiconductors from First Principles
Get PDFCharge transport and dynamics in semiconductors determine the limits of contemporary high-performance electronic devices. Previously, in order to understand the microscopic mechanisms underlying charge transport, and to efficiently find novel materials for new applications, computational methods were limited to using parameterized scattering rates and simplistic band structure models as inputs. However, with ab-initio methods, only the atomic identities and lattice vectors are needed as inputs. These methods have the capability of providing insights not possible with methods that rely on empirical data, and predicting properties for not-yet-synthesized materials.
While ab-initio computation of low-field transport properties have become common in recent years, these methods have not been extensively applied to non-equilibrium phenomena. In addition, the ab-initio simulation of fluctuational properties (such as the diffusion coefficient or power spectral density of current fluctuations) is an area that has been minimally explored. In order to approach quantum-limited noise levels in devices, a better understanding of the mechanisms that govern electronic noise away from equilibrium is needed.
Thus, motivated by this, the overarching goal of this work is to develop and use first-principles methods to gain insight into the scattering processes that govern high-field electronic transport and noise in well-known semiconductors, and to use the same approach to make predictions and identify promising device applications for novel materials.
The warm electron tensor is a quantity that describes the quadratic change of conductivity with electric field, which provides a quantitative way to examine the heating of the electron gas. However, this has not been examined from first-principles previously. In this work, we report the warm electron tensor of n-Si computed over a large temperature range, and find that the most commonly used order of perturbation theory only captures the qualitative change of the warm electron tensor with angle. However, by including the next-to-leading order two-phonon scattering term in our approach, we find near-quantitative agreement. This finding indicates that two-phonon scattering has a non-negligible role to play in transport in nonpolar semiconductors.
We continue our investigation of n-Si by examining the diffusion coefficient and its anisotropy by applying our Boltzmann transport framework to fluctuational variables. We find that the qualitative features of the anisotropy are correct, but its magnitude is greatly underestimated in comparison to experimental data, while the onset of the noise is overestimated. While this suggests an incorrect description of f-type scattering in our work, by computing the frequency dependence of the diffusion coefficient as well as the piezoresistivity (two observables sensitive to the balance of f- and g-type scattering), we find that the qualitative agreement of these two observables with experiment shows that such a discrepancy cannot be due to an incorrect description. Instead, we suggest that the experiment contains charge transport phenomena not accounted for by our electron-phonon scattering framework.
Finally, we use the same approach to investigate the high-field transport and noise in the novel ultra-wide-bandgap semiconductor cubic boron nitride (c-BN). While c-BN is known for its excellent mechanical and thermal properties, its high predicted saturation velocity and breakdown field make it a promising candidate in high-power and high-frequency devices. However, very few experimental and theoretical studies have probed its transport properties. Here, we show that c-BN exhibits a negative differential resistance (NDR) region below 140 K, and show that the cause is due to an abrupt valley repopulation effect with applied electric field. We also show that the intervalley time in c-BN is extremely large, on the order of diamond, and that this large intervalley time causes a distinct noise peak, most prominent at low temperatures. We discuss how the NDR region and large intervalley time make c-BN a potential candidate for transferred-electron devices and Gunn oscillators, respectively.</p
Linear and Non-Linear Interactions Involving Large-Scale Structures in Turbulence
Get PDFThis thesis performs a linear resolvent analysis (McKeon and Sharma, 2010), and a novel quantitative non-linear analysis of the triadic interactions, to study the largescale structures in wall-bounded turbulence.
First, resolvent analysis is applied to a flow over spanwise periodic roughness, to model the large-scale counter-rotating rolls. The experimental data (Wangsawijaya et al., 2020) is utilized to examine both the predictive and data compression capabilities of the resolvent. The improvements by the inclusion of an eddy viscosity and a crude boundary geometry model are also demonstrated. Standard resolvent is able to qualitatively predict the shape of the counter-rotating rolls. The inclusion of eddy viscosity improves the quantitative predictions and combined with the boundary geometry model is able to efficiently represent the data with small differences using only a fraction of the degree of freedom.
Next, we developed a novel framework to quantitatively analyze the triadic non-linear contributions in a turbulent channel. We incorporated the linear resolvent operator to provide the missing link from energy transfer between modes to the effect on the spectral turbulent kinetic energy. The coefficients highlight the importance of interactions involving large-scale structures, for both the large and small-scale forcing and response, providing a natural connection to the modeling assumptions of the quasi-linear (QL) and generalized quasi-linear (GQL) analyses. Specifically, it is revealed that QL and GQL are efficiently capturing important triadic interactions in the flow, and the inclusion of small amounts of wavenumbers into the GQL large-scale base flow quickly captures most of the important triadic interactions.
Finally, by performing spatio-temporal analyses of the triadic contributions to a single mode, we demonstrated the spatio-temporal nature of the triadic interactions and the effect of the resolvent operator. It is shown that the energetic triadic interactions are concentrated in temporal frequencies around a plane where all three wavespeeds are the same, allowing for a truncation of the important triadic interactions. We also demonstrated the linear amplification mechanism of the resolvent, allowing certain triadic interactions to generate a stronger response even with a weak forcing, underscoring the different perspectives offered by the inclusion of the linear resolvent operator into the analyses of the non-linear triadic interactions.</p
Multiscale Design, Fabrication, and Mechanical Analysis of Structural Hierarchies in Functional Materials
Get PDFHierarchical structuring has emerged as a powerful strategy in functional material design to enhance mechanical performance and impart functional properties across multiple scales. In architected materials, leveraging tessellated multiscale geometrical features enables unconventional properties such as ultra-low density and high energy absorption. Similarly, in functional polymers, rational design of molecular chemistries and polymer microstructures allows for tunable mechanical properties and stimuli-responsive behaviors. However, a substantial knowledge gap persists in understanding how multiscale interactions connect to determine the macroscale performance of these materials. This gap arises from challenges in scalable fabrication, multiscale characterization, and limited mechanistic insight from theory and simulations. To address these challenges, this thesis presents a comprehensive approach that integrates scalable fabrication, multiscale characterization, and theoretical modeling to develop hierarchical materials with tunable functionalities. Specifically, we (1) demonstrate scalable fabrication of hierarchical materials using additive manufacturing, (2) investigate the bulk mechanical responses by tuning the smallest level in hierarchical design, and (3) perform multiscale studies to bridge the gap between unit-level interactions and macroscale performance.
In the first study, we explore the role of structural hierarchies in architected polymeric materials for enhanced energy dissipation. Using metasurface-based holographic lithography, we fabricate nano-architected polymeric sheets and demonstrate how geometrical parameters for unit cell design such as relative density and beam aspect ratio influence stiffness, energy dissipation, and deformation modes. These findings highlight the significance of hierarchical structuring in enhancing mechanical performance and establish design principles for scalable manufacturing. In the second study, we focus on dynamic polymers and examine how dynamic crosslinking at the molecular level influences macroscale material responses. A single-step stereolithography approach is developed to tune molecular-level controls in the material. Through multiscale modeling and experimental characterizations, we reveal how dynamic bonding mechanisms govern stiffness, stretchability, and fracture energy. The results underscore the significance of multiscale interactions in tuning mechanical behavior and suggest a pathway for designing materials with programmable responses. In the final study, we build on these insights by integrating molecular-level controls with nonlinear structural responses such as buckling and shape transformations. The central premise is that molecular interactions dictate local responsiveness, while structural geometries can amplify or suppress these responses through localized deformation or stress redistribution. As a demonstration, we explore how tailored viscoelasticity and controlled instabilities can determine the buckling mode of a structural beam. This synergistic interplay highlights the potential of the materials requiring programmed reconfigurability, shape morphing, and stimuli-responsive properties.
The findings presented in this thesis offer a robust framework for bridging molecular-level design with macro-scale performance through scalable fabrication and characterization strategies. By expanding the design space of material-level behavior, this work lays the groundwork for developing next-generation materials with enhanced functionality, adaptability, and intelligence for applications such as soft robotics, healthcare, and sustainable materials.</p