Carrie A. Manore, Ph.D.

Mathematics and Ecosystem Informatics Alumna
Group Leader, Modeling and Observations for Earth Systems (EES-14)
Earth and Environmental Sciences Division
Los Alamos National Laboratory

Biography

Carrie is an Applied Mathematician and Group Leader at Los Alamos National Laboratory who specializes in modeling human, natural, and engineered systems and infectious disease spread. Her work ranges from theory and model analysis to data-driven predictions. She has modeled human, plant, livestock, and wildlife disease across the world in collaboration with dozens of scientists from a broad range of disciplines and national/international organizations.

Building My Career at a National Laboratory

The path to my current position wasn’t direct — or even intentional at first. Certainly building a career at a national laboratory was not on my radar when I started graduate school. My journey after OSU traversed five states and several positions before landing here at Los Alamos National Laboratory, including a postdoc in the Center for Computational Science at Tulane, an NSF postdoc, a staff scientist in the Information Systems and Modeling Group, a deputy group leader in the Theoretical Biology Group, and my current role as group leader (similar to department chair) in Modeling and Observations for Earth Systems Group.

Along the way, I encountered important mentors and developed lessons and mindsets that inspired me and enabled success across a variety of positions. My strategy for success has evolved over time, and I will share a flavor of what it’s like to work at a national laboratory, as well as the skills necessary to advance and grow in our careers.

Finally, I will briefly highlight some of my research and how AI and advanced computing are being integrated into that work.

Valerie Brewer, Integrative Biology
(Advisor: Jamie Cornelius)

Impacts of capture-handling on behavior and fledging success in western bluebirds

Capture-handling is an important and prevalent method of monitoring animal populations, yet it may be perceived by the animal as a predation attempt. Breeding animals must balance their investment in their survival with their investment in their offspring. Predation pressure, even simulated, can have a strong effect on animal behaviors and the optimization of this balance. We examined the effect of capture-handling on territoriality and fledging success in western bluebirds (Sialia mexicana). Here, we followed bluebirds breeding in a nestbox system in Benton County, OR measuring the number of offspring fledged and the breeding male’s response to a simulated territorial intrusion. Males reduced territorial behaviors four days after capture, but captured pairs did not show reduced reproductive success in the current breeding attempt. This work contributes to our understanding of human monitoring impacts on avian species and how animals balance investment in self vs. offspring in the context of predation.

Jingtian Yu, Statistics
(Advisor: Sharmodeep Bhattacharyya)

Applications of statistics in social science

This presentation discusses several applications of statistical methods in social science, with examples from education research, political science, and network analysis. It begins with the LaCuKnoS project, an NSF-supported initiative focused on supporting educators and multilingual learners in science classrooms, and highlights findings on the relationship between language confusion and students’ career preferences. It then considers political science applications, showing how statistical visualization can be used to examine patterns of polarization over time, particularly in attitudes toward combinations of multiple policies. The presentation concludes with an introduction to network data, illustrating how relationships in systems such as transportation, political blogs, research collaboration, and protein interaction networks can be represented and analyzed using graph-based methods. Together, these examples demonstrate the versatility of statistics in transforming complex social and relational data into interpretable evidence for research and decision-making.

Praveeni Mathangadeera, Mathematics
(Advisor: Malgorzata Peszynska)

Coupled snow-soil model

We model heat conduction in snow and soil, both nonlinear, and coupled across an interface. We discuss two approaches: (i) a fully coupled system of nonlinear PDEs for both domains, and (ii) a lumped snow model, a single algebraic snow model resulting in a nonlinear Robin-type boundary condition for the soil. Both models involved over n=13 environmental parameters including the albedo of surface, effects of radiative transfer, and more. For (i), we consider a domain decomposition strategy, with interface conditions to ensure the continuity of temperature and the conservation of heat flux. For (ii), we consider efficient Neural Network parametrizations which can be applied in a decoupled or fully coupled model. Our analyses and simulations contribute to the understanding of the response of the soils in the Arctic to the weather data and help to assess the reliability of the model.

Christian Cunningham, Physics
(Advisor: Bo Sun)

The potential of cellular membrane analysis in breast cancer diagnosis and prognosis

Triple-negative breast cancer (TNBC) accounts for roughly 15% of all breast cancer cases and drives significantly poorer prognosis due to faster spread and fewer therapeutic options. Current biopsy analyses typically focus on nuclei or rely on hematoxylin and eosin (H&E) staining, with additional immunohistochemical (IHC) stains for specific markers. Our work examines how membrane structures and nuclei shapes uncover clinically relevant information. Using a pan-cadherin IHC stain to identify cell membranes alongside nuclei, we extract each cell’s morphology as a quantifiable feature. We then analyze these features with an attention-based model to predict TNBC presence, clinical stage, and pathology grade. The inclusion of both membrane and nuclear data improves prediction accuracy, establishing membrane topology as a tangible, quantifiable biomarker to enhance precision oncology workflows.

Hannah Long, Biochemistry & Biophysics
(Advisor: Sarah Clark)

Bridging the gap: Insights into BLTP-mediated bulk lipid transport

Eukaryotes have incredibly complex membranes that must be maintained for cells to function. One way cells do this is through lipid transfer proteins that shuttle lipids between different cell membranes. Our lab is particularly interested in bridge-like lipid transfer proteins (BLTPs), which are large molecular tunnels that bridge two membranes and transfer lipids in bulk between them. BLTPs are critical for neurological health, and mutations are linked to several severe neurological disorders. Despite their importance, the link between mutations and disease remains unclear because we lack structural information that can help explain how mutations affect protein function. Our lab aims to fill this gap, and we solved the first cryoEM structure of the endogenous BLTP1 complex isolated from C. elegans. This structure provides key insight into binding partner interactions and lipid coordination, laying the foundation to better understand the mechanism of BLTP-mediated lipid transfer and its role in disease.

Kevin Rice, Microbiology
(Advisor: Maude David)

A microbiome-derived milk metabolite improves autism-like behavior and restores wild-type potassium conductances in CNTNAP2 mice

Altered gut-brain communication has been implicated in neurodevelopmental disorders, yet mechanisms linking microbial metabolites to cortical physiology remain unclear. We identified the unsaturated medium-chain fatty acid 5-dodecanoate (5D) as depleted in ASD-diagnosed children relative to non-ASD siblings, with levels positively correlated to milk consumption and milk-derived medium-chain fatty acids. In the CNTNAP2 knockout mouse model of ASD, dietary 5D supplementation reduced autism-like behaviors. Patch-clamp recordings from layer 2/3 prefrontal cortex neurons revealed elevated putative voltage-gated K^⁺ _current density, shorter action potential half-widths, and faster repolarization in CNTNAP2 mice compared to BL6 controls. Notably, 5D supplementation restored these electrophysiological properties to wild-type levels. Using the Simulator of Human Intestinal Microbial Ecosystem (SHIME), we found that 5D abundance increased linearly with palmitate supplementation, supporting a microbial origin. These findings identify 5D as a microbiome-derived metabolite that can normalize cortical excitability and ASD-relevant behaviors.

Karlie Bach, Chemistry
(Advisor: May Nyman)

Tetraperoxometalates for direct air capture of carbon dioxide

While advances in green energy are slowing the rate of new CO₂ emissions, they do nothing to address the CO₂ already accumulated in our atmosphere, trapping heat and pushing us toward catastrophic climate change. Designing materials that selectively capture CO₂ from ambient air is challenging, as other atmospheric gases often compete for the same active sites. Tetraperoxometalates ([M(O₂)₄]^x−; M = Ti, V, Nb) show exciting promise for this challenge. Their reactive peroxide bonds selectively capture CO₂, even at trace atmospheric concentrations, and signal successful capture through a vivid color change, providing a built-in visual indicator of the material’s reactivity. The choice of transition metal center affects the overall performance, with [Ti(O₂)₄]⁴⁻ having a capacity of 7.84 mmol CO₂/g, [V(O₂)₄]³⁻ of 4.01 mmol CO₂/g, and [Nb(O₂)₄]³⁻ of 2.80 mmol CO₂/g. Tetraperoxotitanates exceeded the capacity of current direct air capture technologies, which an average capacity of 3.39 mmol CO₂/g of sorbent.

Ushasi Datta, Chemistry
(Advisor: Marilyn Mackiewicz)

Revolutionizing OCT Imaging: Developing High-Reflectance Nanoparticle OCT Contrast Agents for Targeting Stem Cells

Optical coherence tomography (OCT) offers non-invasive, real-time imaging with micron-scale resolution, yet its clinical and biological potential is constrained by the lack of molecularly specific contrast agents. Overcoming this limitation is critical for advancing OCT in stem cell tracking and regenerative medicine. Here, we introduce hybrid lipid-membrane-coated metal nanoparticles (HMNPs) engineered as targeted OCT contrast agents. By integrating silver and gold nanoparticle cores within a stabilizing lipid membrane, HMNPs achieve high scattering efficiency and tunable localized surface plasmon resonance in the near-infrared (NIR-I and NIR-II) windows. Functionalization with cell-penetrating peptides enables selective uptake by retinal pigment epithelial cells while preserving cellular viability and function. The nanoparticles were optimized to maximize reflectance at ~840 nm, a wavelength matched to current OCT systems. We systematically evaluated how nanoparticle size, morphology, metal composition, and membrane stability govern optical performance, cellular uptake, and biocompatibility. OCT signal enhancement was quantified across nanoparticles. Our results demonstrate robust OCT contrast enhancement without aggregation-driven signal loss or cytotoxicity, establishing HMNPs as a versatile platform for molecularly informed OCT imaging. This work expands OCT beyond structural imaging and enables its application to stem cell monitoring and regenerative therapies.

Genevive Sheehan, Chemistry
(Advisor: Marilyn Mackiewicz)

Unlocking Axonal Transport: Nanoparticle-based Tracer for Glaucoma Detection

Current methods for glaucoma diagnosis focus on detecting vision loss via perimetry and on assessing retinal ganglion cell (RGC) loss via optical coherence tomography (OCT). However, significant morphological changes in RGCs, including alterations in the axonal cytoskeleton, mitochondria, and dendrites, can be observed via histopathology and electron microscopy (EM) even before these changes become detectable. Studies utilizing fluorophore-labeled cholera toxin subunit B (CTB) have shown that these early axonal changes correspond with measurable deficits in axonal transport, a crucial function of RGCs. Despite this, EM has yet to establish a direct link between ultrastructural breakdown and transport deficits at the subcellular level. We hypothesize that CTB-conjugated silver nanoparticles can be used to study axonal transport across multiple imaging modalities and scales. Using click chemistry, we developed biocompatible hybrid lipid-coated silver nanoparticles (HMNPs) conjugated with Oregon Green dyes and CTB proteins. The CTB's conjugation to the hybrid lipid coating was confirmed using thin-layer chromatography, NMR, and fluorescence. Our findings indicate that RGC cells can be successfully targeted with CTB-conjugated HMNPs with minimal inflammation when injected intravitreally. Additionally, these nanoparticles allow for real-time tracking of axonal transport in vivo using multiple imaging techniques, contributing to our understanding of axonal cytoskeletal breakdown and transport failure in glaucoma.

MJ Strike, Integrative Biology
(Advisor: Sarah Henkel)

The Power of Waves: Marine Renewable Energy Technology

As offshore energy industries expand worldwide, new habitats are created for marine life. Oregon’s active wave energy climate is particularly suitable for wave energy development, but the addition of new substrates to the environment may have unintended consequences for the region’s ecological dynamics. Biofouling organisms (e.g., mussels, barnacles, and bryozoans), potentially including invasive species, may use wave energy buoys as habitat. Buoys may also alter recruitment dynamics as larvae from rocky reefs may use them as stepping stones. Several models of offshore wave energy buoys are presented, the likes of which could be deployed at the PacWave energy test site in Oregon in the near future. In addition, the audience can engage with a mini wave tank containing two model buoys, one of which is "biofouled" with "barnacles" (beads) and "algae" (fabric), demonstrating the impact of biofouling on buoy movement and energy-harvesting efficiency.

Zejing Wang, Mathematics
(Advisor: Chad Giusti)

Introduction To Persistent Homology Transform

Applied topology has shown surprising power in recent AI research. In particular, persistent homology provides a refined algebraic structure that can be extracted from both discrete and continuous data via computational methods, while it has a theoretic background from algebraic topology.

Casey Rummelhart, Chemistry
(Advisor: Addison Desnoyer)

Tuning Metal Nuclearity in Ni-Al Complexes for the Reduction of CO2

As a key greenhouse gas, carbon dioxide (CO2) is largely responsible for observed global warming and its predicted progression. Considering atmospheric CO2 levels continue to climb by approximately 2 ppm per year, the development of effective methods to transform this gas into valuable chemical feedstocks remains a pressing challenge within the synthetic chemistry community. Because CO2 readily reacts with strong nucleophiles, hydride-based nucleophiles derived from abundant hydrogen gas (H2) will aid in sustainable transformations. These hydride nucleophiles will be generated using Frustrated Lewis pairs, the combination of a Lewis acid and Lewis base, unable to form a classical Lewis adduct. This frustration is proven to heterolytically cleave hydrogen into a proton (H+) and hydride (H-), together allowing CO2 to be transformed into formic acid. As a precursor to the cleaner fuel source methanol, approximately 1 million tonnes of formic acid are produced annually, making this process industrially valuable.

Kelsi Ramos, Chemistry
(Advisor: Alison Bain)

The Optical Properties of Aqueous Carbonate-Containing Aerosol

Esther Julius, Chemistry
(Advisor: May Nyman)

Facile Synthesis of Organotin Oxo Clusters

Organotin oxo clusters are of important interest due to their structural tunability and stability, making them useful for nanolithography applications in the production of integrated circuits. However, their synthesis often involves complex procedures and special apparatus. In this work, we explored a simple and efficient resin-assisted method for the synthesis of [C36H34O18Sn6],[C8 H27 O22Sn12 ]2+and Sn20 clusters from Butyltin trichloride (BuSnCl₃) under mild conditions at room temperature. The anion exchange resin enabled controlled hydrolysis of the precursor, promoting the self-assembly of these cluster topologies. Reaction parameters, including solvent and anion type, were systematically investigated to understand their effect on cluster formation and film uniformity. The resulting products were isolated and characterized using electrospray ionization mass spectrometry (ESI-MS), small-angle X-ray scattering (SAXS), single-crystal X-ray diffraction (SCXRD),and nuclear magnetic resonance (NMR) spectroscopy. This approach provides a straightforward and reproducible route to multiple tin oxo cluster sizes and offers a more efficient synthesis of alkyltin molecules for nanolithography.

Hannah Long, Biochemistry & Biophysics
(Advisor: Sarah Clark)

Bridging the gap with insights into LPD3-mediated bulk-lipid transport

Bridge-like lipid transfer proteins (BLTPs) are large protein complexes that transfer lipids in bulk between two organelle membranes at membrane contact sites. There are five BLTP family members conserved from yeast to humans, all of which are critical for neurological health. Our lab focuses on elucidating the structures of native BLTPs to better understand their function and the role they play in neurological health. We solved the first cryoEM structure of native C. elegans BLTP1 (also called LPD3), revealing key insights into complex architecture and the mechanism of lipid transfer.

Hannah Stuwe, Biochemistry & Biophysics
(Advisor: Elisar Barbar)

Local SR Structure and Variant Mutations Fine-tune LRH-Mediated Oligomerization of SARS-CoV-2 Nucleocapsid Protein

Anishika Nagar, Chemistry
(Advisor: Marilyn Mackiewicz)

Hybrid Lipid-Shielded Silver Nanotriangles Enable Shape Stability and Subcellular Targeting

Nanoparticle probes in subcellular imaging often lack targeting specificity, become trapped in endo-lysosomes, and perturb organelle function. To address these challenges, we design hybrid lipid-shielded silver nanotriangles that directly overcome these barriers, achieving programmable, shape-stable mitochondrial targeting while preserving plasmonic functionality and biocompatibility. This strategy enables precise, non-disruptive subcellular imaging and functional studies. Quantitative imaging reveals selective mitochondrial localization and demonstrates that targeting efficiency depends on intracellular trafficking pathways rather than on uptake alone. Notably, this probe enables intact mitochondrial labeling with preserved function, distinguishing it from approaches that may compromise organelle integrity. By tuning surface ligand density, we establish a noninvasive labeling regime that retains both targeting fidelity and mitochondrial integrity. These findings provide a mechanistic understanding of nanoparticle-based mitochondrial targeting by directly linking surface chemistry to intracellular behavior and biological outcomes. Overall, this platform offers a generalizable strategy for creating stable nanoparticle probes with improved targeting accuracy and functional performance for subcellular imaging.

Reginald Appiah-Kubi, Biochemistry & Biophysics
(Advisor: David Hendrix)

Comparison of long-read assembly and population genomics reveals patterns of genetic diversity among Y chromosomes in hops (Humulus lupulus L.)

Hallee Boyd, Chemistry
(Advisor: May Nyman)

Counterion Driven Formation of Lanthanide Polyoxomolybdate Structures

Mitchell Kim-Fu, Chemistry
(Advisor: Jennifer Field)

Classroom Demonstration Used for Visualization of Three Environmental Partition Coefficients

In environmental chemistry and toxicology courses, chemical partitioning is a fundamental concept needed for predicting the exposure, fate, and transport of chemicals. However, visualizing and understanding partitioning behavior can be challenging in classroom settings. This study describes a simple, hands-on activity using commercially available food-grade menthol crystals and vegetable oil, as a safer proxy for octanol, to demonstrate air-water, octanol-water, and octanol-air partition coefficients in a single-beaker setup. Students simultaneously calculate partitioning constants from the public database PubChem and place their olfactory experience in the context of their calculated value. This activity was implemented in a combined undergraduate and graduate student toxicology and received unanimously positive feedback. Students reported increased confidence in using chemical databases and a clearer conceptual understanding of partitioning in environmental contexts. Additionally, integrating a demonstration into the classroom setting increased student participation and discussion on atmospheric transport and deposition.

Emily Werner, Chemistry
(Advisor: Alison Bain)

Nonionic surfactants affect water activity of binary aqueous solutions

Surfactants have been found in environmentally relevant aqueous solutions like aerosol droplets. These surfactants can impact aerosol water activity and, therefore, hygroscopic growth. The Aerosol Inorganic-Organic Mixtures Functional groups Activity Coefficients (AIOMFAC) is a group contribution model that predicts the water activity of solutions containing inorganic ions and organic species. One shortcoming of AIOMFAC is that predictions for water activity when strong surfactants are present in solutions have not been validated against experimental measurements. Here, AIOMFAC model predictions for water activity of solutions containing strong surfactants are compared to our experimental results. Water activity for surfactants that have a polyethylene glycol (PEG)-like tail, is better predicted by AIOMFAC when using an oxyethylene subgroup rather than ether and alkyl subgroups. Our model-measurement validation demonstrates the importance of appropriate subgroup selection when using AIOMFAC for water activity predictions.

Ayantika Dan, Chemistry
(Advisor: Dipankar Koley)

Novel Carbon-Based Soft Hydrophobic Electrodes for Selective Detection of Metabolites using Electrochemical Techniques

Electrochemical sensing technology is advancing towards flexible electrodes. Most of these electrodes exhibit issues with successful electrodeposition and mechanical deformities, resulting in reduced electron transfer efficiency. There is a need to overcome the limitations of flexible conductive and chemically modified electrodes and to increase their robustness and reliability for real-world applications. Our lab's newly developed carbon-based soft electrodes are designed for enhanced flexibility, improved electron transport, and enhanced surface modification capabilities. These electrodes have a hydrophobic surface. Hydrophobic metabolites, namely Uric acid, are chosen as a model system to prove this unique surface property of the electrode. These analytes are crucial biomarkers for gout, diabetes, cardiac, and neural diseases in humans. Most electrodes need surface modifications to detect these molecules. The newly developed electrodes address the need for surface modification and do not require it to detect hydrophobic molecules. It has good sensitivity and can potentially detect these biomarkers from real-world samples. These findings indicate that the newly developed electrodes provide a platform for non-invasive biosensing applications, including wearable sensor technologies and clinical diagnostics.

Samararathne Muhandiramge Sachindee Chandula, Chemistry
(Advisor: Dipankar Koley)

Bioanalytical Characterization of Metabolic Modulation in Multi-Species Oral Biofilms by Zn-BAG Resin Composites

Oral diseases have a significant impact on human health worldwide, affecting individuals across all age groups. These conditions often challenge the durability and performance of restorative materials used in dentistry. Traditional antimicrobial dental materials are designed to kill bacteria directly, but this approach can disturb the natural microbial balance. This study focuses on metabolic modulation rather than microbial eradication. By incorporating zinc-doped bioactive glass into resin composites, we can achieve sustained zinc release that influences biofilm metabolism and promotes beneficial microbial activity. The study emphasizes bioanalytical monitoring of the biofilm–biomaterial interface using highly sensitive, non-invasive sensors. Miniaturized impedance-based wire sensors allowed real-time tracking of biofilm growth and volume during the growth period, while potentiometric pH sensors measured localized pH dynamics at the interface. This research highlights a new approach in dental material design that actively modulates the microbial ecosystem rather than simply killing bacteria. By integrating electrochemical sensing and real-time bioanalytical measurements, this study establishes a framework for developing advanced dental materials that support a healthier, caries-antagonistic biofilm phenotype.

Emmanuel Oguadimma, Mathematics
(Advisor: Nathan Gibson)

Multidimensional Structure-Preserving FDTD Methods for Maxwell's Equations in Kerr-Debye-Lorentz Media

We present a structure-preserving finite difference time domain (FDTD) method for the three-dimensional Maxwell--Kerr--Debye--Lorentz system. Built on the classical Yee staggered-grid discretization of Maxwell's equations and auxiliary differential equations for the Debye and Lorentz material responses, the method is designed to preserve key structural properties of the continuous model. In particular, a modified exponential update for the Kerr susceptibility handles the stiffness in the nonlinear material response, preserves nonnegativity, and is asymptotic-preserving in the Kerr relaxation limit, recovering the instantaneous Kerr--Lorentz model. We prove a discrete energy identity under a CFL stability condition and exact preservation of a discrete divergence constraint in the source-free case. The resulting nonlinear electric-field update is implemented efficiently through Newton--GMRES with matrix-free Jacobian--vector products, making the method practical for large-scale multidimensional simulations. Numerical experiments confirm second-order accuracy, verify the divergence-preserving property, and show that the scheme captures nonlinear optical phenomena including soliton propagation and harmonic generation. Overall, the proposed method provides a robust and efficient multidimensional FDTD scheme for nonlinear dispersive Maxwell models, with potential applications in computational photonics including simulation-based optimization.

Sultana Parvin Rumi, Chemistry
(Advisor: Addison Desnoyer)

A Tunable Multinucleating Ligand Platform for Cooperative Small-Molecule Activation

Tepary Cooley, Biochemistry & Biophysics
(Advisor: Maria Purice)

A glial neuropeptide regulates organismal oxidative stress

Abraham Kpirikai, Biochemistry & Biophysics
(Advisor: Nathan Mortimer)

The Drosophila PDGF/VEGF signaling pathway regulates host immunometabolism in response to parasitoid infection.

Xinyu 'Erica' Li, Statistics
(Advisor: Lan Xue)

Cluster analysis of functional data with random effects using p-spline and pairwise comparisons

Measurement error from device inaccuracy and improper use can substantially degrade the performance of functional data clustering. We propose a clustering method that accounts for complex measurement error by incorporating subject level random effects to capture within-subject variability. The proposed method approximates individual trajectories with a truncated power basis and exploits the connection between penalized splines and the linear mixed model to introduce random effects into the smoothing penalty representation. Clusters are formed via a pairwise comparison penalty imposed on fixed effect coefficients. Theoretically, we establish the convergence rate of the proposed method for estimating the true mean functions. We evaluate the proposed method through simulation studies and apply it to energy expenditure data from a childhood obesity study. The results show that the proposed method improves the estimation of group mean functions and achieves higher clustering accuracy than naive approaches.

Davis Sharts, Biochemistry & Biophysics
(Advisor: Richard Cooley)

Specificity from Infidelity: Biophysical Insights Toward an Improved phospho-Threonine Genetic Code Expansion System

Cells use phosphorylation of serine and threonine residues to regulate protein function, fate, and signaling pathways. Despite this central role, only 3% of reported phosphosites have assigned function, highlighting a need to provide functional characterization of the phospho-proteome. Genetic code expansion (GCE) is a unique tool allowing site-specific incorporation of non-canonical amino acids (ncAAs), including phospho-serine (Sep), into recombinantly expressed proteins in E. coli using orthogonal tRNA/tRNA-synthetase (RS) pairs from archaea. A GCE system exists for phospho-threonine (pThr), however it is significantly less efficient, and misencodes other phosphorylated amino acids. To address these issues, we solved the structure of the pThr tRNA synthetase (pThrRS) to 1.78Å and performed biophysical characterization. Comparison to the SepRS reveals distinct active-site architecture underlying substrate recognition and inefficiency, along with a shift in oligomerization equilibria. These findings provide a framework for engineering improved pThrRS variants and more efficient GCE systems for phosphoproteome studies.

Anthony Winchell, Physics
(Advisor: Oksana Ostroverkhova)

Novel characterization of spin waves in twisted CrSBr/CrSBr homostructures via DFT/QFT modeling and experiments

We present a combined experimental, density functional theory (DFT), and quantum field theory (QFT) approach to understanding spin waves in twisted 2D magnetic heterojunctions. In particular, we investigate twisted CrSBr/CrSBr bilayers to reveal how interlayer twist angles and structural variations affect spin-wave generation, propagation, and coherence. Experimentally, we fabricate twisted CrSBr homostructures with controlled rotation angles and probe their steady-state and time-resolved optical responses as functions of twist angle, magnetic field, and temperature. Complementary DFT calculations provide the local atomic geometry and exchange coupling parameters, which are incorporated into a QFT framework to model spin-wave dispersion relations in twisted and defect-engineered structures. By correlating theoretical dispersion features with experimentally mapped spin-wave dynamics, this integrated methodology reveals how interlayer twist and nanoscale morphology affect magnon behavior in van der Waals magnets.

Oluwaseun Adu, Integrative Biology
(Advisor: Michael Blouin)

Who Wins and Why? Uncovering the Genomic Basis of Resistance Polymorphism in a Snail–Parasite System.

Evan Flint, Statistics
(Advisor: Robert Trangucci & Lisa Madsen)

Bayesian Estimation of Cumulative Survival Rates in Cormack-Jolly-Seber Models

Cormack-Jolly-Seber models are used to estimate individuals' survival probabilities in a population between repeated recapture events. Often the cumulative survival probability across multiple recaptures is of more interest than survival probabilities between each successive recapture. Cumulative survival can be estimated by maximum likelihood, as implemented by software including Mark. However, maximum likelihood frequently encounters computational challenges and inflated variance estimates in settings where point estimates are near the parameter space boundary. In contrast, Bayesian estimation avoids these challenges but suffers from highly informative priors under parameterizations generally used in practice. We propose a Bayesian approach with independent gamma priors on the negative log of interval survival parameters. We show that under this method, point and interval estimates of cumulative survival are substantially improved relative to commonly used prior structures and maximum likelihood. We illustrate our findings with an application to Snake River Chinook Salmon, the animal population motivating this study.

Kamrul Chowdury, Mathematics
(Advisor: Nathan Gibson)

Uncertainty Quantification in the Isotropic Cold Plasma Model

In plasma modeling, the collision time may vary because of uncertainty in material properties or surrounding conditions. To capture this effect, we treat the collision time as a random parameter in the isotropic cold plasma model. Using polynomial chaos (PC) expansion, the random model is converted into a coupled deterministic system. We then discretize the resulting PC system with the Kashiwa--Fukai (KF) finite-difference time-domain (FDTD) scheme. This poster presents results on stability, convergence, and dispersion, showing the accuracy and reliability of the method for random cold plasma simulations.

Victory Obieke, Mathematics
(Advisor: Vrushali Bokil)

Computational Modeling of Spatial Soliton Propagation and Scattering

Nolan Herron, Biochemistry & Biophysics
(Advisor: Juan Vanegas)

Decoding Teixobactin–Lipid II Binding: How Local Environments and Hydration Shape Interaction Strength

Teixobactin is a promising antibacterial agent with potent bactericidal activity, particularly against antibiotic-resistant bacterial strains. Despite extensive studies on teixobactin binding to lipid II, it remains unclear how the antibiotic achieves such potency given that its primary target is the pyrophosphate and sugar moiety of lipid II. An important mechanistic question is how does teixobactin avoid non-specific interactions with other molecules that contain phosphates or pyrophosphates within the cell? To address this, we employed well-tempered metadynamics simulations with PLUMED and GROMACS to construct two-dimensional free energy surfaces of teixobactin interactions with lipid II and other peptidoglycan precursors in both solution and membrane environments. We find that hydration of the pyrophosphate group on lipid II plays a key role in defining the affinity of teixobactin for its lipid target. The reduced solvent accessibility of the lipid II headgroup in the membrane environment results in the formation of a tightly bound teixobactin complex that sits in a deep energy well. In contrast to this, interaction of the lipid II headgroup with teixobactin in solution leads to frequent binding and unbinding events with shallow minima and a low energy barrier for association/dissociation. These findings provide mechanistic insights into teixobactin’s mode of selectivity and highlight how membrane-associated hydration environments enable stable binding to lipid II while minimizing nonspecific interactions.

Sima Ziyaee, Chemistry
(Advisor: Claudia Maier)

Single-Cell Detection of Mitochondria-Targeted Compounds by Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry

Andriana Zourou, Biochemistry & Biophysics
(Advisor: Myriam Cotten)

Seeing Red: Extraction of Bioactive Peptides from Devaleraea mollis

Emilee Lance, Microbiology
(Advisor: Ryan Mueller)

Proteomic Stable Isotope Probing as a Precision Tool for Identifying Gut Microbial Fiber Utilization in a Complex Community

Cameron Call, Biochemistry & Biophysics
(Advisor: Alysia Vrialas-Mortimer)

Aggregation of Lamin Protein in Aging and Disease

Nicholas Bretz, Biochemistry & Biophysics
(Advisor: Nathan Mortimer)

Assessing the nature of venom vesicles: A parasitoids strategy for hostile takeover

Courtney Clement, Microbiology
(Advisor: Ryan Mueller)

Impacts of marine heatwaves on the eelgrass (Zostera marina) phyllosphere microbiome

Seagrasses act as ecosystem engineers, providing valuable habitat for juvenile fish, preventing erosion, and sequestering carbon. Marine heatwaves (MHWs), which are increasing in both frequency and intensity globally, are a major threat to seagrass health and are associated with shoot loss and increased disease. To ascertain the effects of MHWs on seagrass phyllosphere microbiomes, Zostera marina shoots gathered from Yaquina Bay, OR were exposed to a simulated MHW (+5C) in outdoor mesocosm tanks. Leaf samples were collected during the heat treatment, and three times during a recovery period spanning 29 days. Over the course of the 17-day simulated MHW, the phyllosphere experienced shifts in the microbial community structure that persisted into the recovery period. Additionally, the heat community rate of change accelerated during the MHW period, before slowing to control levels during the recovery period. MHWs appear to destabilize the seagrass phyllopshere microbiome and accelerate vectors of microbial change.

Bryce Pettit Estell, Microbiology
(Advisor: Martin Schuster)

Quorum sensing in Pseudomonas aeruginosa as a bistable population switch

Laila Brubaker, Microbiology
(Advisor: Stephen Atkinson)

An Ex Vivo Assay for Characterizing Chemical Inhibition of Myxozoan Spore Attachment to the Fish Host

Gabriel Park, Statistics
(Advisor: Claudio Fuentes)

Fault Detection and Signal Realignment in Malfunctioning or Perturbed Data

Tara Conrad, Microbiology
(Advisor: Kimberly Halsey)

Fourier Transform Infrared Spectroscopy: A Rapid, Spectroscopic Method to Characterize the Macromolecular Composition of Natural Phytoplankton Communities

Phytoplankton macromolecular composition varies according to taxonomy and physiology. Macromolecular composition reflects caloric content and is therefore an indicator of cell nutritional quality. Studies on macromolecular composition of natural phytoplankton communities are rare because of biomass limitations but could improve our understanding of how phytoplankton nutritional value drives marine food webs and the subsequent carbon flux from the surface to the deep ocean. Fourier-Transform Infrared spectroscopy (FT-IR), which interrogates the absorbance of infrared radiation by characteristic molecular bonds, offers a rapid, low-cost technique to assess changes in coarse macromolecules (e.g., protein, lipid, carbohydrates). Although FT-IR has been applied in culture-based studies, to our knowledge it has yet to be applied to characterize phytoplankton macromolecular composition in field samples. Here, we report FT-IR-based macromolecular characterization of natural phytoplankton communities across diverse spatial and time scales. Results will yield new information about phytoplankton physiology and their impacts on trophic transfer efficiency.

Sarah Perkel, Biochemistry & Biophysics
(Advisor: Sarah Clark)

Isolation and Analysis of Native Primate ASIC1a from Brain Tissue

Ion channels play essential roles in human physiology by depolarizing a cell’s membrane potential in response to extracellular signals. Acid Sensing Ion Channels (ASICs), which form multimeric complexes, play important roles in diverse processes from blood pressure regulation to pain sensation. However, despite the importance of ASICs to human health, major questions remain about their regulation, complex make-up, and channel architecture in vivo. We have developed a native protein isolation method to specifically enrich ASIC1a-containing complexes and analyzed this population with mass spectrometry (MS) analysis. We found that ASIC1a-complexes contain numerous other ASIC isoforms, suggesting the native population contains heteromeric complexes, as well as numerous potential interacting proteins.

Sneha George, Chemistry
(Advisor: Dipankar Koley)

A Micro-Gap Model for Real-Time Monitoring of Biofilm Metabolism and Localized Corrosion at the Tooth–Restoration Interface

Tinu Anthony, Chemistry
(Advisor: Dipankar Koley)

Disposable Prussian Blue Sensor for the Detection of Picomoles of Hydrogen peroxide from Multispecies Dental plaque Biofilm

Hydrogen peroxide (H₂O₂) is a reactive oxygen species with dual roles in biological and environmental systems: at low concentrations it functions as a signaling molecule in cellular processes and microbial interactions, while at higher levels it induces oxidative stress and tissue damage. In the oral microbiome, small amounts of H₂O₂ from commensal bacteria help suppress pathogens and maintain balance. Accurate detection at very low concentrations is therefore essential for understanding microbial dynamics, host–microbe interactions, early diagnostics, and monitoring oxidative stress. Conventional methods often lack sufficient sensitivity, selectivity, or practicality for detecting low micromolar or sub-micromolar levels in complex matrices.

This work presents a Prussian Blue (PB)-modified H₂O₂ sensor integrated with a microvolume electrochemical cell for sub-micromolar detection. PB enables selective detection at low overpotential, minimizing interference, while the hydrogel-based microvolume cell amplifies signals to detect picomole levels. Using a sensitive coulometric method, the system shows linear response from 0.5–5 nmol with a 250 pmol detection limit. It demonstrates high reproducibility and effective detection in complex samples such as dental plaque biofilms with minimal preparation, offering a practical platform for ultrasensitive H₂O₂ sensing.

Nadia Gonzalez, Biochemistry & Biophysics
(Advisor: Alysia Vrailas-Mortimer)

Context-Dependent Interactions Between Copper Homeostasis and Parkinson’s Disease

Hao Yue, Chemistry
(Advisor: Marilyn Mackiewicz)

Smart silver nanoplates cloaked in hybrid lipids and armed with antibodies bring precision, stability, and biocompatibility to imaging triple-negative breast cancer

Emily Hiatt, Chemistry
(Advisor: May Nyman)

Simplifying Capture: Using Zirconium Hexamers for Removal of PFAs from Water

Mahya Payazdan, Biochemistry & Biophysics
(Advisor: Adrian Gombart)

Toll-like receptor-mediated repression of vitamin D-induced cathelicidin antimicrobial peptide gene expression may involve NF κB

Antimicrobial peptides (AMPs), are small peptides, that play significant role in the innate immunity through defense against pathogens and immune modulation. In Mammals AMPs are classified into Cathelicidins and Defensins. LL-37 is the active form of Cathelicidin and is produced from the hCAP18 precursor through proteolytic cleavage and is expressed by immune and epithelial cells. Vitamin D, particularly its active form 1,25(OH)₂D₃, strongly induces LL-37 expression via vitamin D receptor-mediated transcription. However, pathogens have evolved ways to suppress LL-37 by interfering with transcription factors, histone modifications, or membrane repulsion strategies. Our preliminary data shows that, Toll Like Receptors (TLRs) activation, especially TLR4 pathway, represses vitamin D-induced CAMP expression, but not expression of another vitamin D-target gene like CD14 in PMA-differentiated THP1. The underlying mechanisms of this suppression are not fully understood. This study aims to uncover mechanisms regulating suppression of vitamin D-induced CAMP expression by employing CRISPR-Cas9 technology to knock out key genes involved in TLR signaling pathways such as IRF3, MYD88, TRIF, TLR4 and NFKB1 using THP1 cell lines, the research seeks to identify genes that suppress CAMP expression and assess their role in immune response to infections.

Pavel Sengupta, Chemistry
(Advisor: Dipankar Koley)

Biocompatible soft carbon-based ion-selective electrodes for the point-of-care quantification of Calcium in human saliva

Quantification of Calcium is predominantly performed through an invasive blood testing. Ionized Calcium testing is prevalent through ion-selective electrodes though it requires strict sample pre-treatment methods, immediate analysis and does not provide fast measurements. Calcium levels can inform onset of kidney disease, certain kinds of cancers, or parathyroid gland abnormalities. Abnormalities in Calcium content can affect heart rate, induce muscle spasms or osteoporosis. There is a need to find better non-invasive ways of measuring Calcium from biofluids. In this work, we provide a portable device framework to provide point-of-care quantification of ionized Calcium in human saliva using soft biocompatible carbon-based Calcium ion-selective electrodes. Our sensors provide selective detection of Calcium with a wide linear range of 1 μM – 0.1 M and low detection limit of 5 μM. We switch to a non-invasive biofluid – saliva and provide a portable device to measure Calcium from microliter volumes of raw saliva.

Allissa Van Steenis, Microbiology
(Advisor: Maude David)

Enrichment of enteroendocrine cells using magnetic activated cell sorting to investigate microbial taxa in ASD and neurotypical populations

Autism spectrum disorder (ASD) includes persistent deficits in social communication and restricted, repetitive behaviors. In the United States, prevalence is about 3.2% (one in 31 children). ASD frequently co-occurs with anxiety and gastrointestinal disorders. Animal studies suggest mechanisms such as increased gut permeability, inflammation, and altered enteroendocrine cell (EEC) distribution. While considerable efforts have been made to uncover the relationship between dysbioses of the gut microbiome in autism, and in spite of growing evidence correlating perturbations of the EECs with ASD, the mechanisms triggered by these interactions and the specific microbial taxa involved remain unknown. We will isolate EECs from human duodenum using negative selection. Non-EEC epithelial cells will be removed via biotinylated antibodies targeting alkaline phosphatase (ALPI). Neutravidin magnetic beads will capture labeled cells, enriching EECs. Crosslinking will then identify bacteria associated with these cells. Subsequent sequencing will characterize taxa and potential functional interactions underlying gut–brain signaling alterations in ASD and reveal candidate microbial drivers for future mechanistic validation studies

Hayden Vaughn, Statistics
(Advisor: Tate Jacobson)

An Additive Model Importance Method

Caroline Hernández, Microbiology
(Advisor: Maude David)

Identifying Microbial Effectors of Neuropod-Mediated Gut-Brain Communication through Whole-Cell Crosslinking

Dani Gaudette, Biochemistry & Biophysics
(Advisor: Dan Liefwalker)

Taming the Seahorse: Wrangling MYC-Driven Cancer Metabolism

Oluwasen Adu, Integrative Biology
(Advisor: Michael Blouin)

Genome Wide Association Study of Biomphalaria glabrata snail and it's Schistosome Parasite

Schistosomiasis is a parasitic disease of humans that is caused by trematodes in the genus Schistosoma. Schistosomes require aquatic snails as an intermediate host to complete their lifecycle. Infected snails shed parasites into water, which then infect humans through their skin. With WHO’s goal to eliminate schistosomiasis as a global health problem by 2030, a renewed emphasis on snail control was enacted. Because different genomic regions are important in different snail-parasite population combinations, it is important to look at many combinations of Biomphalaria glabrata and Schistosoma mansoni populations to find more regions. Via genome-wide association studies (GWAS), genomic regions associated with variation in resistance of Biomphalaria glabrata population to infection by Schistosoma mansoni population will be identified. The goal is to identify new snail genomic regions or previously known regions that could confer resistance to schistosome infection. This work might point to new strategies for blocking transmission of schistosome parasites.

Vera Alenicheva, Chemistry
(Advisor: Vincent Remcho)

A Microfluidic Paper-Based Assay for the Quantification of CBD and THC

Rapid and accurate quantification of cannabinoids, including cannabidiol (CBD) and tetrahydrocannabinol (THC), is crucial for medical, legal, and regulatory applications. Differentiating these cannabinoids is particularly important due to their contrasting properties: CBD is non-psychoactive, while THC is psychoactive. This study introduces a novel microfluidic paper-based assay designed for point-of-care diagnostics, addressing the demand for portable, efficient testing methods. Using selective laser ablation, a paper-based device was developed featuring circular test zones (8.5 mm) optimized for UV-Vis spectroscopy, fluorescence analysis, and RGB-based colorimetric detection, which utilizes color changes captured through the Red, Green, and Blue channels for analyte evaluation. Cannabinoid quantification is achieved through a reaction with Fast Blue BB Salt (FBBB) under alkaline conditions, forming distinctive chromophores—orange for CBD and pink for THC—with a maximum absorption at 480 nm used for both chromophores, alongside a fluorophore specific to THC, with excitation at 480 nm and emission at 655 nm. ImageJ software enables analysis of RGB values extracted from the wells of the captured image of the array-based paper device, while optimized reagent concentrations ensure low detection limits and robust linearity. Bench-top UV-Vis and fluorescence analysis further validate the device's reliability. For all assays, 0.5M NaOH provided optimal conditions, enabling smoother color formation and higher signal intensity. THC fluorescence analysis achieved a 0.77 µg/mL limit of detection (LOD) at 10 mM FBBB, as higher concentrations quenched fluorophore formation. For THC colorimetric detection, 15 mM FBBB was optimal, with an LOD of 24.87 µg/mL. The CBD assay performed best with 10 mM FBBB, achieving an LOD of 48.45 µg/mL. The assay demonstrates significant advantages over traditional methods, offering rapid, accessible testing for field and clinical applications. Future work will focus on stabilizing FBBB with polymer supports, developing cannabinoid extraction protocols for biological samples, and enhancing detection limits with alternative substrates. Integration into a prototype device aims to expand point-of-care capabilities, establishing this portable assay as a transformative tool for on-site cannabinoid testing across diverse environments.

Lucas Allan, Chemistry
(Advisor: Tim Zuehlsdorff)

FC2DES: Modeling 2D Electronic Spectroscopy for Harmonic Hamiltonians

Franck-Condon spectroscopy has long been a ubiquitous method for modeling the linear optical properties of molecular chromophores. Its working principal, which requires the nuclear motion of the molecule to be represented as harmonic, leads to a robust approximation of the optical properties of the system and is freely accessible to researchers through standard computational chemistry software packages such as Gaussian. It is thus desirable to extend this Franck-Condon methodology to modern, more complex non-linear experiments. However, there previously were no Franck-Condon formulations for non-linear experiments that could be applied to systems of molecular scale. In my work, I present the FC2DES approach, providing the first, readily accessible formulation for the Franck-Condon simulation of non-linear spectroscopic experiments for systems of molecular size.

Hallee Boyd, Chemistry
(Advisor: May Nyman)

Characterization of Trivalent Lanthanide Keggin Phosphomolybdate Sandwich Clusters

Lanthanide polyoxomolybdate (PMo) sandwich complexes offer a versatile platform for examining actinide chemistry by using non-radioactive surrogates to elucidate structural and chemical trends. By exploring molybdate-based systems, comparisons can be extended to tungsten analogues offering insights into how variations of the POM framework influence coordination behavior, solution behavior across the lanthanide series. Single Crystal X-Ray Diffraction reveals coordination geometries, metal–ligand bond lengths, and overall complex bonding angles that reflect periodic trends across the lanthanide series based on ionic radius. By investigating the complexes’ behavior in the solution state via 31P NMR and X-Ray scattering, the distribution of monomeric (1:1) and dimeric (2:1) POM:lanthanide species are examined. The chemical tunability, radiation resistance, high solubility, and compatibility of PMos with a wide range of characterization techniques support their use in microgram-scale studies, enabling new opportunities to investigate possible trends relevant to actinides while minimizing the challenges associated with handling radioactive materials.

Daniel Malone Buoy , Statistics
(Advisor: Claudio Fuentes, Sarah Emerson)

Representative Sampling Methods for K-Fold Cross Validation

K-fold cross validation is widely used for assessing the accuracy of predicative regression and classification models. We investigate improving representativeness of the training set and testing set for K-fold cross validation in the context of random forests, and in particular exploring the potential for using a mixture of stratified sampling and cluster sampling to reduce variance in estimation of model performance measured by residual standard error and apparent error rate. Simulations and applications to real-world data are considered to compare performance to competing methods for cross validation, namely to the standard random fold assignment.

Olivia Burleigh, Integrative Biology
(Advisor: Virginia Weis)

Transcription Factor-targeted ChIP-Seq for Smad3-mediated TGF-β Signaling in Heat-stressed Aiptasia

Reef-building corals rely on a crucial symbiosis with endosymbiotic dinoflagellates, yet the cellular and molecular mechanisms behind this relationship remain unclear. Recent studies suggest the host’s innate immune system plays a key role in establishing, maintaining, and breaking down the symbiosis. One signaling molecule involved is TGF-β, which promotes immune tolerance and symbiosis establishment. However, the downstream signaling pathways, particularly those leading to immunosuppression, are not fully understood. To investigate this, I adapted chromatin immunoprecipitation sequencing (ChIP-Seq) protocols for use with the Smad3 transcription factor in Exaiptasia diaphana—a model for coral symbiosis—where ChIP-Seq has only been applied to histones before. This approach enables genome-wide analysis of Smad3-DNA interactions and will help reveal how gene regulation differs depending on the symbiotic state or under acute heat stress, shedding light on the transcriptional mechanisms that support or disrupt symbiosis in response to environmental change.

Jun Cai, Integrative Biology
(Advisor: Virginia Weis)

Effect of Sphingolipid Metabolic Pathway Inhibition and Knockdown on Cnidarian-Algal Symbiosis

Climate change has accelerated coral bleaching and the destruction of coral reefs, threatening the survival of these crucial hotspots of ecological, cultural, and economic services for millions worldwide. Despite the threat of increased bleaching events, many aspects of the cellular mechanisms of coral-algal symbiosis remain under-studied. A greater understanding of these fundamental aspects of coral biology is essential for predicting how corals respond to ongoing environmental change and for developing successful conservation strategies. Sphingolipids are important lipid signalers that regulate many cellular processes related to host-microbe interactions. We investigated the functional role of the sphingolipid pathway during symbiosis establishment in Exaiptasia diaphana (commonly known as Aiptasia), a model system for coral-algal symbiosis studies. We used chemical inhibitors for two key sphingolipid enzymes during symbiont uptake and used RNAi to knock down the expression of these enzymes to assess their importance in symbiosis. We examine the effects of the inhibitors and RNAi on aposymbiotic, symbiotic, and recolonized polyps. Changes in symbiont density and photosynthetic efficiency were quantified, physiological changes were visualized with microscopy, and qPCR was performed to confirm the effects of gene knockdown.

Giovanni Crestani, Integrative Biology
(Advisor: Molly Burke)

Genomics of Experimentally Evolved Postponed Reproduction in Drosophila Melanogaster

Evolve and resequence (E&R) experiments with model organisms can reveal the genetics underlying complex traits and adaptation dynamics. We sequenced genomes of Drosophila melanogaster populations evolved for postponed reproduction over 20 generations. These populations, maintained on a 70-day cycle, evolved distinct life-history phenotypes, including increased lifespan, compared to controls. Comparing populations with the same selection treatment but different evolutionary histories allowed us to test how past evolution shapes future adaptation. Our results demonstrate that allele frequencies align with selection treatment rather than recent evolutionary history, indicating evolutionary convergence within selection regimes. We also show that adaptation for postponed reproduction has a polygenic basis and occurs rapidly through shifts in standing genetic variation.

Ushasi Datta, Chemistry
(Advisor: Marilyn Mackiewicz)

Unveiling the Hidden Properties: How Nanomaterial Surface Chemistry and Biomimetic Systems Shape Reflectance and Contrast

Nanomaterials have great potential to enhance optical coherence tomography (OCT) imaging through their reflectance properties. However, understanding how their surface chemistry and physicochemical features—such as NP surface coatings, environmental pH, and salinity—impact reflectance remains a challenge. Most nanomaterials developed for OCT imaging function in the NIR-I window (650–950 nm) and exhibit smaller scattering cross sections in the NIR-II window (1000–1700 nm). Gold nanoparticles (AuNPs) and silver nanoparticles (AgNPs), with strong localized plasmon resonance (LSPR) bands, are promising candidates that can be tuned to match the operating wavelength of optical coherence tomography (OCT). However, there is a gap in knowledge regarding how the surface chemistry of nanomaterials, their biological exposure conditions, and environmental factors alter their reflectance properties and OCT contrast properties. Here we examine how varying physiological salt concentrations, pH levels, and biological media, including gelatin, intralipid, and melanin, change the reflectance behavior of nanomaterials. Gelatin mimics extracellular tissue, intralipid replicates retinal tissue, and melanin regulates light while protecting the retina. Hence, it is essential to understand how biological tissue mimics fluorescence reflectance. This knowledge is necessary for identifying nanomaterials for use as OCT contrast agents and will be beneficial to the biomedical imaging community.

Konstantin Drallios, Chemistry
(Advisor: Thomas Osborn Popp)

3D Printable Radiofrequency Coils

To perform nuclear magnetic resonance (NMR) spectroscopy, a perpendicular magnetic field must be applied to the sample. Experimentally, this condition is met by placing the sample in a radiofrequency (RF) coil. The fabrication of RF coils is generally performed using wire, but the range of possible coil geometries that can be wrapped from a strand of wire are limited, and it can be difficult to ensure dimensional accuracy and reproducibility during winding. Here, we present a new approach that integrates computer-aided design (CAD), 3D printing, and electrochemical copper plating to produce RF coils. By using a home-built Python program to compute 3D magnetic field distributions for CAD-modeled coils, we further enable our fabrication technique by optimization and prediction of coil magnetic field. As a validation we apply this technique to generate a solenoid coil with optimized magnetic field homogeneity and demonstrate its capabilities to perform NMR spectroscopy.

Rudranil Dutta, Chemistry
(Advisor: Claudia Maier)

Identification and Quantitation of Bioactive Alkaloids in Withania Somnifera

Withania somnifera (WS) has been widely used as a medicinal herb for a variety of health problems related to stress and aging. WS is rich in withanolides, withanolide glycosides, tropane alkaloids, and alkyl ferulates. Withanolides and withanolide glycosides have been traditionally associated with the bioactivity and health promoting effects of WS extracts. However, we have recently discovered that the tropane alkaloids found in WS extracts are a potential new class of bioactive compounds in WS. However, high resolution tandem mass spectrometry, mass spectral analysis, authentic standard synthesis and NMR verification are all required to elucidate alkaloid compositional differences in WS. Finally, developing a fast and reliable method for quantifying tropane alkaloids in plant material and commercial products is forthcoming using multiple reaction monitoring mass spectrometry to quantify these alkaloids reliably.

Arpa Ebrahimi, Chemistry
(Advisor: Claudia Maier)

Characterizing the Lipidomic and Proteomic Profile of the 5xFAD Alzheimer’s Disease Mouse Model: A Comparative Study Using MALDI Imaging Mass Spectrometry

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by amyloid-beta (Aβ) plaque accumulation, disrupting neuronal communication and leading to cognitive decline. Although plaque deposition is a hallmark of AD, early neurodegeneration can occur before visible plaques form. Using the 5xFAD mouse model, we investigated lipidomic alterations in the hippocampus of young (1 month old) and old (8–9 months old) 5xFAD mice compared to wild-type controls. Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging enabled spatial visualization of lipid distribution within the hippocampus. Our analysis revealed distinct changes in specific lipid classes, particularly phospholipids and sphingolipids, in young 5xFAD mice prior to detectable plaque formation. Comparative profiling across age groups highlighted progressive lipid dysregulation coinciding with disease development. These findings suggest that lipid metabolism alterations contribute to early AD pathology and may serve as potential biomarkers for pre-symptomatic detection. This study also establishes an optimized MALDI imaging workflow for future applications in brain lipidomics research.

Jessica Etter, Chemistry
(Advisor: Claudia Maier)

LC-QTOF and LC-TIMS-qQTOF MS Analysis of Fecal Inoculum Biotransformation Products and LC-MRM-MS Analysis of Human Withanolide Pharmacokinetics of an Ashwagandha Supplement

Withania somnifera (ashwagandha) and its withanolide constituents have been shown to exhibit a variety of effects in humans, including stress relief, anti-inflammatory activity, and neuroprotective properties. Despite ashwagandha’s growing popularity, there remains a significant lack of research into the metabolism of withanolides, particularly within the human gut microbiome. Methods include the use of in silico metabolism prediction software, in vitro fermentation using fecal inocula, and pharmacokinetic studies. This comprehensive approach contributes to understanding the interplay between gut microbiota and bioactive, bioavailable metabolites derived from ashwagandha, aiding both pharmacological development and insights into the role of gut health in these processes. Preliminary results indicate a variety of in vitro transformations of withanolides following incubation with gut microbiota from human fecal inocula. Additionally, data suggest that the bioactive compound sominone may be generated in vivo from precursor molecules found in ashwagandha, potentially through transformation in the gastrointestinal tract.

Caroline Hernandez, Microbiology
(Advisor: Maude David)

Whole-Cell Crosslinking Reveals Direct Lactobacillaceae and Rhizobiaceae Interactions with Host Duodenal Neuropods

Communication along the microbiota-gut-brain axis occurs through neuronal, endocrine, metabolic, and immune pathways. While most research has focused on indirect signaling via microbial metabolites and hormone secretion, a direct afferent pathway was recently discovered between neuropods—specialized enteroendocrine cells (EECs)—and vagal afferents. Neuropods, in contact with gut contents, may transmit signals to the brain within milliseconds. However, direct receptor-ligand interactions between microbiota and neuropods remain unexplored. In this study, we used Sulfo-SBED, a cell-impermeable crosslinker, to covalently capture direct interactions between murine duodenal EECs and gut microbes. 16S amplicon sequencing, metagenomics, and single-cell proteomics revealed 24 enriched bacterial taxa—exclusively from Lactobacillaceae, Rhizobiaceae, and Pseudomonadaceae—interacting with EECs. Notably, Ligilactobacillus animalis, L. murinus, and Agrobacterium spp. were consistently enriched. Proteomic analysis identified five EEC-associated proteins, including pro-glucagon and ghrelin, as significantly enriched in crosslinked samples. This work provides the first evidence of direct microbial interactions with neuropods, offering novel insight into gut-brain communication.

Esteban Hernandez, Chemistry
(Advisor: Jennifer Field)

Experimental pKa Values of Substituted and Unsubstituted Perfluoroalkyl Sulfonamides via 19F NMR

Perfluoroalkyl sulfonamide-containing substances (FASAs), such as perfluorooctane sulfonamide (FOSA) and perfluorobutane sulfonamide (FBSA), are a class of per- and polyfluoroalkyl substances (PFAS) found throughout the environment in water, soil, and organisms. Exposure to FASAs leads to negative health effects like nephrotoxicity and hepatotoxicity. Despite their prevalence and toxicity, relatively little information exists on how FASAs partition among phases in the environment. One characteristic of a molecules that determines its transport through the environment is the acid dissociation constant, pKa. The pKa informs when a molecule is predominantly charged or neutral, which has implications on partition coefficients, aqueous solubility, sorption to sediment, and partitioning into biota. In this work, five FASAs were analyzed by 19F NMR to determine experimental pKa and the effect of chain length and substitution on the nitrogen on the pKa.

Lucas Kolanz, Physics
(Advisor: Davide Lazzati)

Cosmic Dust Bunnies

Cosmic dust is an important part of many astrophysical processes, however, its formation mechanism in the early universe is still unknown. We use Soft Sphere Discrete Element Method simulations of dust formation to study the structures that arise in an astrophysical environment consistent with a supernova remnant. We find that temperature plays an important role in the resulting structure of coalescing dust grains.

Weiqi ‘Grace’ Li, Statistics
(Advisor: Yuan Jiang)

Reframing Spatial Transcriptomics Prediction: From Regression to Classification

Spatial transcriptomics (ST) is an emerging technology that sequences gene expression from intact tissue while preserving spatial context, promising unprecedented insights into tissue structure and disease mechanisms. However, its high cost and technical complexity hinder widespread clinical adoption. To address this, recent methods leverage deep learning to predict spatial gene expression from widely available histology (H&E) images. Although these models demonstrate modest correlation with true expression levels, their ability to distinguish between expressed and non-expressed genes, or between low and high expression, remains limited (AUC ≈ 0.6). This project is, to our knowledge, the first to reframe the prediction task as classification rather than regression. Applied to real-world, subcellular-resolution ST data, our classification-based model significantly improves performance—achieving AUCs of approximately 0.8 (binary decision) and 0.9 (probability-based). This significant boost in discriminative power provides a foundation for broader applications of gene expression reconstruction in disease screening and treatment outcome prediction.

Sarah Louie, Biochemistry & Biophysics
(Advisor: Richard Cooley/Ryan Mehl)

Optimizing genetic code expansion technology to access post-translationally modified proteins

Genetic code expansion (GCE) enables the incorporation of non-canonical amino acids (ncAAs) into proteins at UAG stop codons using an orthogonal tRNA–aminoacyl-tRNA synthetase (aaRS) pair. This allows site-specific chemical modifications for studying protein structure, function, and interactions. A major application of GCE is encoding post-translational modifications (PTMs) such as phosphorylation. Protein phosphorylation on serine (pSer) and threonine (pThr) regulates key processes including cell growth and signaling. While efficient GCE systems exist for pSer, current pThr systems lack the efficiency to produce functional proteins. To improve phospho-threonine encoding, we are modifying key GCE components. We screened orthologs of the L-Thr kinase pduX to increase intracellular pThr availability. Additionally, we are evolving a more efficient pThr aaRS by generating a mutation library targeting its active site. Enhancing these components will improve system fidelity and accessibility, enabling better tools to study the role of pThr in protein and cellular function.

Praveeni Mathangadeera, Mathematics
(Advisor: Malgorzata Peszynska)

Computational Modeling of the Nonlinear Heat Equation in Frozen Soil and Snow

Heat conduction in snow is modeled by a nonlinear parabolic partial differential equation which modifies the Stefan problem of liquid/ice phase transition with a free boundary. In soil, heat conduction is a nonlinear parabolic equation which is smoother than the Stefan problem. We discuss and compare three time stepping formulations for discretization of these models: the fully implicit, sequential and semi-implicit approaches. For the soil problem we also compare the use of different primary unknowns. Next, we explore the pointwise snow model which solves a single algebraic equation instead of a PDE and thus approximates the snow PDE model. This pointwise model involves n=13 environmental parameters including the albedo of surface. We analyze the sensitivity of the solutions and set up a ML regression model: that allow us to assess the robustness of our computational model as well as to understand the uncertainty associated with the parameters and the model itself. Our simulations and analyses contribute to the understanding of the response of the soils in the Arctic to the weather data and help to assess the reliability of the model.

Anshika Nagar, Chemistry
(Advisor: Marilyn Mackiewicz)

Shielded Nanoparticles: Advancing X-Ray Fluorescence Microscopy with Oxidant-Resistant Nickel and Cobalt

Synchrotron-based X-ray fluorescence microscopy (XFM) is a powerful tool for imaging the distribution of transition metals in biological specimens, yielding insights into metal-related physiological and pathological processes. However, conventional organic fluorophores are unsuitable for XFM due to their inability to emit detectable X-rays, rapid degradation, and non-specific mapping. Alternative methods like Fourier Transform Infrared Microscopy (FTIRM) and immunofluorescence show promise, but ptychography with XFM faces challenges in identifying organelles. We propose using functionalized cobalt (CoNPs) and nickel (NiNPs) nanoparticles to enhance cellular visualization in XFM scans. Their distinct K-edges enable effective metal mapping without interference from other biological metals. Nevertheless, these nanoparticles must overcome issues such as surface oxidation and aggregation. Our presentation will focus on strategies to stabilize NiNPs and CoNPs, improving their water solubility and preventing oxidation. Ultimately, these nanoprobes aim to facilitate the mapping of metal ion pools in cells and clarify their roles in diseases.

Luke Nearhood, Physics
(Advisor: Patti Hamerski)

Computing the Tension

Computational Physics is a key part of the landscape of 21st century physics. However there has been limited research on effective teaching techniques in computational physics. Especially in upper division lab courses, where practicing physics can look very different to introductory courses, as students join the physics community and develop their physics identity. We conducted interviews of students in a one-credit junior level computational physics lab course at a large public research university. We analyzed these interviews through the framework of Activity Theory. In so doing we sought to identify tensions in the activities of the course and thus where improvements might be made, and to understand the general dynamics and character of the course as it currently exists. We identified tensions between group work expectations, building proficiency at computing, and other aspects of the course. The specific contextual features of this computational physics lab add nuance to our findings.

Joline Nguyen, Biochemistry & Biophysics
(Advisor: Sarah Clark)

Isolation of Scarce Membrane Protein Complexes from C. elegans

Membrane proteins serve critical roles in many biological processes and have been identified as potential targets for the development of therapeutic treatments. Biochemical and structural studies of many membrane proteins have been hindered by challenges associated with recombinant expression in heterologous cell lines and low abundance in native tissue. To overcome these challenges, we have developed a strategy that enables isolation of sufficient amounts of endogenous membrane proteins from transgenic Caenorhabditis elegans for biochemical, biophysical, and structural analyses. We detail a step-by-step workflow from large-scale growth of genetically modified C. elegans to protein purification utilizing standard laboratory equipment. We also describe how a protein’s quantity and quality can be assessed in small scale via fluorescent detection size exclusion (FSEC) analysis, and conditions for protein isolation in small-scale experiments can be adjusted and optimized to improve the yield and purity of the protein of interest. To demonstrate the feasibility and utility of this strategy, we describe herein a successful isolation of a low-abundance, multi-subunit “protein complex A” that cannot be reconstituted in cell culture. Our isolation strategy can help facilitate the studies of challenging, biomedically relevant membrane proteins.

Victory Chiamaka Obieke, Mathematics
(Advisor: Vrushali Bokil)

Compatible Energy Preserving Discretizations for Nonlinear Optical Wave Propagation: The Maxwell-Duffing Approach

This poster presents the modeling and numerical simulation of electromagnetic wave propagation in nonlinear optical media using the Maxwell-Duffing model. The material’s nonlinear response is described by a cubic Duffing oscillator coupled with Maxwell’s equations in a non-magnetic, non-conductive setting. We derive the associated energy relations for the one-dimensional system to guide the development of stable numerical methods. A high-order spatial discretization is introduced using a fully discrete leap-frog finite-difference time-domain (FDTD) scheme combined with operator splitting. Emphasis is placed on second-order time and higher-order space accuracy for resolving traveling wave solutions. We establish the energy stability of the proposed Yee-type FDTD schemes and support our theoretical findings with numerical experiments. This work addresses key challenges in simulating nonlinear photonic phenomena and provides robust tools for advancing research in nonlinear wave propagation.

Emily Palmer, Statistics
(Advisor: Yuan Jiang)

A Group Penalization Framework for Detecting Time-Lagged Microbiota-Host Associations

We present a framework to identify time-lagged associations between abundances of longitudinally sampled microbiota and a stationary response (final health outcome, disease status, etc). We introduce a definition of the time lag by imposing a particular grouping structure on the association pattern of longitudinal microbial measurements. Using group regularization methods, we identify these time- lagged associations including their strengths, signs, and timespans. Simulations accurately identify both time lags and signal strengths. We apply this framework to find specific gut microbial taxa and their lagged effects associated with increased parasite worm burden in zebrafish.

Madison Phelps, Mathematics
(Advisor: Malgorzata Peszynska)

Nonlinear Solvers in Permafrost Applications

We study the nonlinear heat equation with phase change between liquid and ice where the energy to change phase is given by the enthalpy-temperature relationship, a(q). For Stefan problems, a(q) is usually a multivalued graph where the phase transition resembles a shifted Heaviside function. For permafrost problems, a(q) is continuous but its inverse requires a nonlinear solver. For soil with trapped air and large pores, we propose a model denoted (R*) where a(q) is a multivalued graph with piecewise continuous properties. We show how to extend analytical solutions for the Stefan problem to (R*). Also, the robustness of our numerical algorithm will be shown, where we use various local nonlinear solvers such as Newton-Anderson, to study the performance and convergence of our solver. We also describe progress towards a coupled Salinity model.

Kevin Rice, Microbiology
(Advisor: Maude David, Kenton Hokanson)

Electrical Characterization of Primary Enteroendocrine Cells: Developing Tools to Screen Novel Microbial Neuroactive Compounds

The vagus nerve plays a central role in gut-brain communication, relaying microbial signals sensed by enteroendocrine cells (EECs) to the brainstem. A subset of EECs, known as neuropods, synapse onto vagal afferents and release cholecystokinin (CCK), providing a rapid, direct gut-to-brain connection. However, the electrophysiological properties of CCK-EECs remain poorly characterized. Here, we performed whole-cell patch clamp recordings on primary mouse EECs labeled with CCK-GFP and isolated via FACS. Using a co-culture with vagal nodose neurons, we cultured stable, electrically active EECs and measured intrinsic properties, including resting potential, input resistance, and capacitance. Most cells fired action potentials upon stimulation and exhibited voltage-dependent Nav and Kv currents. This work establishes a foundational characterization of murine CCK-EECs. Additionally, by combining co-culture, whole-cell crosslinking, and single-cell proteomics, we identified microbial taxa and EEC-associated proteins, such as pro-glucagon and ghrelin, that can now be tested for their effects on gut-brain signaling.

Casey Rummelhart, Chemistry
(Advisor: Addison Desnoyer)

Frustrated Lewis Pairs Ligand for the Transformation of Carbon Dioxide to Chemical Feedstocks

With the deleterious role of carbon oxides on climate change, scientists are desperately seeking to discover methods to efficiently sequester and functionalize these gases. While this idea has been explored via frustrated Lewis pairs stoichiometrically, catalytic examples remain scarce. It is proposed that a rationally designed, bimetallic ligand will alleviate these challenges and allow an understanding of the mechanism of transformation. This design makes use of a modular bis-phosphine NacNac system, which will allow for tunability of both electronic and steric parameters. Additionally, this design takes advantage of inverted charge density compared to typical FLP examples, displaying a concentrated site of reactivity within the bimetallic pair.

Pavel Sengupta, Chemistry
(Advisor: Dipankar Koley)

Quantifying Dissolved Oxygen in Biofilms with Non-invasive Flexible Amperometric Oxygen Sensors

Dissolved oxygen (DO) is one of the most frequent and important parameters for measurement in numerous scientific fields such as healthcare, environmental, food industries, energy, to name a few. Traditional oxygen sensors such as the optical, fluorescence, colorimetric and electrochemical oxygen sensors such as galvanic or polarographic types might be bulky, cause chemical leakage or degradation, and lack the flexibility to use in biological systems such as biofilms, human cells and tissues, and in small volume or miniaturized measurements. Furthermore, traditional methods of oxygen detection are often obtained as an end-result quantification and often not performed in real-time. Fabricated a unique flexible amperometric (electrochemical) oxygen sensor which addresses some of the challenges of traditional electrochemical oxygen sensors widely used. A system has been proposed with supported data to measure DO in bacterial specimen. This will be applied in quantifying the oxygen concentrations in oral biofilms non-invasively in real-time.

Michael Sieler, Microbiology
(Advisor: Thomas Sharpton)

Modeling the zebrafish gut microbiome’s resistance and sensitivity to climate change and parasite infection

As global water temperatures rise due to climate change, parasitic infections are expected to spread and intensify. Yet, the gut microbiome may buffer hosts against infection, particularly if it remains stable under environmental stress. Using zebrafish (Danio rerio), we tested whether the gut microbiome is resistant or resilient to elevated temperature, parasite exposure, and their combination—and whether these responses relate to infection outcomes. Adult zebrafish were exposed to the intestinal parasite Pseudocapillaria tomentosa at 28°C, 32°C, or 35°C, and fecal microbiomes were sampled across 42 days. We found that temperature and parasite exposure independently affected gut microbiome diversity, while temperature moderated parasite-microbiome associations. Surprisingly, higher temperatures reduced infection burdens. Our results suggest temperature alters the microbiome’s response to parasitic stress and may mitigate infection severity. This work highlights how climate change can yield unexpected effects on host health.

Gavin Tovar, Statistics
(Advisor: Robert Trangucci, Sarah Emerson)

Sequential Approach to K-Fold Cross-validation---Computational Reduction Technique

Cross-validation (CV) and variants of it are widely used for predictive error estimation and model selection criterion. We investigated using sequentially developed CV folds to reduce computational time, inspired by the common one standard error rule found in the CV literature. Simulations and applications to real world data were used to compare the performance to competing CV methods.

Hao Yue, Chemistry
(Advisor: Marilyn Mackiewicz)

Targeted X-ray Imaging Agents for Visualizing Triple-Negative Breast Cancer

Triple-negative breast cancer (TNBC) is a highly aggressive subtype lacking estrogen, progesterone, and HER2 receptors, limiting treatment to chemotherapy and surgery. Its poor prognosis and resistance to conventional imaging highlight the need for improved diagnostic tools. We developed hybrid-lipid-coated silver nanoparticles (AgNPs) functionalized with anti-mouse IgG secondary antibodies for targeted TNBC imaging. These nanoparticles bind PI-selective primary antibodies that recognize phosphatidylinositol lipids overexpressed on TNBC cells, enabling indirect labeling with enhanced specificity and signal amplification. The lipid coating improves stability and biocompatibility. To assess performance in a tumor-like context, we are generating 3D TNBC spheroids to evaluate nanoparticle uptake, fluorescence distribution, and cytocompatibility, supporting future diagnostic development.

Sima Ziyaee, Chemistry
(Advisor: Claudia Maier)

Exploring Cellular Heterogeneity through Single-cell Proteomics

Single-cell proteomics provides a powerful approach to explore cellular heterogeneity and capture subtle biological changes often masked in bulk analyses. Here, we present two studies that highlight its critical importance. First, MDA-MB-231 breast cancer cell lines stably overexpressing Bcl-2 were generated through electroporation and clonal selection in G418-containing media. High Bcl-2 expression clones (MDA-MB-231/Bcl-2) were identified by Western blotting and subjected to chemical library screening alongside vector controls. Single-cell analysis revealed specific protein expression changes, including differential expression of BCL2L1, that were not detectable at the bulk population level. The second study investigates PC12 cell differentiation into sympathetic ganglion neurons induced by nerve growth factor (NGF). Using single-cell proteomic profiling, we not only monitor differentiation progression and identify protein and peptide markers but also unveil cellular heterogeneity within the differentiated population, which remains obscured in traditional bulk proteomics. Together, these studies underscore the vital role of single-cell analysis in revealing hidden molecular diversity.

Madison Phelps, Mathematics
(Advisor: Malgorzata Peszynska)

Nonlinear Solvers in Permafrost Applications

We study the nonlinear heat equation with phase change between liquid and ice where the energy to change phase is given by the enthalpy-temperature relationship, a(q). For Stefan problems, a(q) is usually a multivalued graph where the phase transition resembles a shifted Heaviside function. For permafrost problems, a(q) is continuous but its inverse requires a nonlinear solver. For soil with trapped air and large pores, we propose a model denoted (R*) where a(q) is a multivalued graph with piecewise continuous properties. We show how to extend analytical solutions for the Stefan problem to (R*). Also, the robustness of our numerical algorithm will be shown, where we use various local nonlinear solvers such as Newton-Anderson, to study the performance and convergence of our solver. We also describe progress towards a coupled Salinity model.

Sarah Louie, Biochemistry & Biophysics (Advisor: Richard Cooley/Ryan Mehl)

Optimizing genetic code expansion technology to access post-translationally modified proteins

Genetic code expansion (GCE) enables the incorporation of non-canonical amino acids (ncAAs) into proteins at UAG stop codons using an orthogonal tRNA–aminoacyl-tRNA synthetase (aaRS) pair. This allows site-specific chemical modifications for studying protein structure, function, and interactions. A major application of GCE is encoding post-translational modifications (PTMs) such as phosphorylation. Protein phosphorylation on serine (pSer) and threonine (pThr) regulates key processes including cell growth and signaling. While efficient GCE systems exist for pSer, current pThr systems lack the efficiency to produce functional proteins. To improve phospho-threonine encoding, we are modifying key GCE components. We screened orthologs of the L-Thr kinase pduX to increase intracellular pThr availability. Additionally, we are evolving a more efficient pThr aaRS by generating a mutation library targeting its active site. Enhancing these components will improve system fidelity and accessibility, enabling better tools to study the role of pThr in protein and cellular function.

Weiqi ‘Grace’ Li, Statistics
(Advisor: Yuan Jiang)

Reframing Spatial Transcriptomics Prediction: From Regression to Classification

Spatial transcriptomics (ST) is an emerging technology that sequences gene expression from intact tissue while preserving spatial context, promising unprecedented insights into tissue structure and disease mechanisms. However, its high cost and technical complexity hinder widespread clinical adoption. To address this, recent methods leverage deep learning to predict spatial gene expression from widely available histology (H&E) images. Although these models demonstrate modest correlation with true expression levels, their ability to distinguish between expressed and non-expressed genes, or between low and high expression, remains limited (AUC ≈ 0.6). This project is, to our knowledge, the first to reframe the prediction task as classification rather than regression. Applied to real-world, subcellular-resolution ST data, our classification-based model significantly improves performance—achieving AUCs of approximately 0.8 (binary decision) and 0.9 (probability-based). This significant boost in discriminative power provides a foundation for broader applications of gene expression reconstruction in disease screening and treatment outcome prediction.

Lucas Kolanz, Physics
(Advisor: Davide Lazzati)

Cosmic Dust Bunnies

Cosmic dust is an important part of many astrophysical processes, however, its formation mechanism in the early universe is still unknown. We use Soft Sphere Discrete Element Method simulations of dust formation to study the structures that arise in an astrophysical environment consistent with a supernova remnant. We find that temperature plays an important role in the resulting structure of coalescing dust grains.

Esteban Hernandez, Chemistry
(Advisor: Jennifer Field)

Experimental pKa Values of Substituted and Unsubstituted Perfluoroalkyl Sulfonamides via 19F NMR

Perfluoroalkyl sulfonamide-containing substances (FASAs), such as perfluorooctane sulfonamide (FOSA) and perfluorobutane sulfonamide (FBSA), are a class of per- and polyfluoroalkyl substances (PFAS) found throughout the environment in water, soil, and organisms. Exposure to FASAs leads to negative health effects like nephrotoxicity and hepatotoxicity. Despite their prevalence and toxicity, relatively little information exists on how FASAs partition among phases in the environment. One characteristic of a molecules that determines its transport through the environment is the acid dissociation constant, pKa. The pKa informs when a molecule is predominantly charged or neutral, which has implications on partition coefficients, aqueous solubility, sorption to sediment, and partitioning into biota. In this work, five FASAs were analyzed by 19F NMR to determine experimental pKa and the effect of chain length and substitution on the nitrogen on the pKa.

Caroline Hernandez, Microbiology
(Advisor: Maude David)

Whole-Cell Crosslinking Reveals Direct Lactobacillaceae and Rhizobiaceae Interactions with Host Duodenal Neuropods

Communication along the microbiota-gut-brain axis occurs through neuronal, endocrine, metabolic, and immune pathways. While most research has focused on indirect signaling via microbial metabolites and hormone secretion, a direct afferent pathway was recently discovered between neuropods—specialized enteroendocrine cells (EECs)—and vagal afferents. Neuropods, in contact with gut contents, may transmit signals to the brain within milliseconds. However, direct receptor-ligand interactions between microbiota and neuropods remain unexplored. In this study, we used Sulfo-SBED, a cell-impermeable crosslinker, to covalently capture direct interactions between murine duodenal EECs and gut microbes. 16S amplicon sequencing, metagenomics, and single-cell proteomics revealed 24 enriched bacterial taxa—exclusively from Lactobacillaceae, Rhizobiaceae, and Pseudomonadaceae—interacting with EECs. Notably, Ligilactobacillus animalis, L. murinus, and Agrobacterium spp. were consistently enriched. Proteomic analysis identified five EEC-associated proteins, including pro-glucagon and ghrelin, as significantly enriched in crosslinked samples. This work provides the first evidence of direct microbial interactions with neuropods, offering novel insight into gut-brain communication.

Giovanni Crestani, Integrative Biology
(Advisor: Molly Burke)

Genomics of Experimentally Evolved Postponed Reproduction in Drosophila Melanogaster

Evolve and resequence (E&R) experiments with model organisms can reveal the genetics underlying complex traits and adaptation dynamics. We sequenced genomes of Drosophila melanogaster populations evolved for postponed reproduction over 20 generations. These populations, maintained on a 70-day cycle, evolved distinct life-history phenotypes, including increased lifespan, compared to controls. Comparing populations with the same selection treatment but different evolutionary histories allowed us to test how past evolution shapes future adaptation. Our results demonstrate that allele frequencies align with selection treatment rather than recent evolutionary history, indicating evolutionary convergence within selection regimes. We also show that adaptation for postponed reproduction has a polygenic basis and occurs rapidly through shifts in standing genetic variation.