University of Georgia

Abstracts for the 2026 CURO Summer Research Fellows are listed in alphabetical order by the student’s name. The CURO Summer Research Fellowship Final Forum features our students’ work with their mentors over summer 2026.

Abstracts from 2025 are available at https://curo.uga.edu/students/summer-research-fellowship/2025-abstracts/.

Our students are:

 

Resolving the Chromatin Architecture of Maize Neocentromeres using Fiber-seq

Nicholas Alexander
Majors: Genetics, Ecology, Biology
Faculty Mentor: Kelly Dawe, Genetics and Plant Biology, Franklin College of Arts and Sciences
Additional Co-Authors: Kempton Bryan, Dong-Won Kim

Complex genetic transformations of maize will be vital in maintaining an efficient and robust food supply system as we travel into increasingly unpredictable environmental conditions. To make effective modifications, a neochromosome could be added to serve as a platform for these changes, avoiding silencing genome effects from cis-regulatory elements. An artificial centromere must be generated on this molecule to make associations with the mitotic spindle for proper segregation during cell division. Centromere growth and movement across cell types is not well understood; therefore, we need a better method of directly measuring centromeres. Fiber-seq is a long-read sequencing method which utilizes an adenine methyltransferase (Hia5) to add methylation markers onto adenines. When this reaction occurs with intact chromatin, these markers can only be added to open regions (euchromatin) in the genome. Using Fiber-seq, we want to measure centromere position through euchromatin depletion patterns; however, we must first obtain a working methodology of this protocol. We first optimized our nuclei isolation protocol for Fiber-seq confirming nuclei quality and methylation with a PCR based assay. From this method we were able to sequence Hia5-treated DNA using Nanopore sequencing and begin bioinformatics analysis through cluster computing. We expect to see similar methylation profiles as seen in a short-read euchromatin sequencing method called ATAC-seq. This work has laid the foundation for us to fully develop Fiber-seq in maize which allows us to analyze artificial centromeres more effectively, giving us better insight in the creation of stable neochromosomes.

Remote Neuropsychological Assessment in Clinical High Risk for Psychosis: Considering Two Digital Platforms

Vivian Austin
Major: Psychology (Neuroscience Emphasis)
Faculty Mentor: Gregory P. Strauss, Psychology, Franklin College of Arts and Sciences
Additional Co-Author: Luyu Zhang

Cognitive assessments are essential tools for understanding neuropsychological function, but traditional in-person testing can present barriers. Remote cognitive assessments have emerged as an increasingly used alternative for healthy populations and individuals with schizophrenia. However, little research has examined their use for individuals at Clinical High-Risk (CHR) for psychosis, despite evidence of small-to-moderate cognitive deficits and recommendations for frequent monitoring. This study examined whether two remote assessment platforms detect cognitive differences between CHR individuals and healthy controls (CN). 26 CHR participants and 31 demographically-matched CN were recruited via advertisements and referrals. Across 7-day spans, participants received six daily prompts to complete one of three mini neuropsychological tasks via the Ecological Momentary Assessment (EMA). At the end of this week, participants completed the TestMyBrain Digital Neuropsychology Toolkit during a 90-minute video call with a researcher. EMA outcomes were analyzed using multilevel regression, and TestMyBrain outcomes using independent-samples t-tests. Results revealed a marginally significant association between CHR status and time to complete EMA tasks (p = .07, d = -.52), with CHR participants completing tasks more slowly. However, neither EMA accuracy scores nor overall composite differed by group (p = .41, d = -.23; p = .12, d = -.43). CHR participants performed significantly worse on the TestMyBrain Toolkit compared to CN (p = .01, d = -.69). Conclusion: TestMyBrain detected cognitive deficits in CHR comparable to those reported for in-person assessments, supporting its potential as a remote neuropsychological tool. Additional research is needed to determine the sensitivity and validity of EMA-based cognitive assessment in CHR.

Mental Health Court Program Outcomes of Those with Externalizing and Internalizing Disorders

Alice Brodsky
Majors: Psychology, Criminal Justice
Faculty Mentor: Orion Mowbray, Ralston Institute for Behavioral Health and Developmental Disabilities, College of Family and Consumer Sciences

Individuals with serious mental illness are disproportionately represented in the criminal justice system, yet traditional probation models often fail to account for how different mental illnesses shape behavior and needs. Mental Health Courts (MHCs) were developed to address this gap by diverting individuals with mental illness from traditional correctional supervision into programs that emphasize treatment, monitoring, and judicial oversight. The primary goal of this study is to examine whether individuals with internalizing disorders and externalizing disorders experience different outcomes from MHC programs. Findings may support the refinement of supervision approaches that are more responsive to participant needs, potentially enhancing long-term rehabilitation and reintegration outcomes. This study used quantitative secondary data collected by the Council of Accountability Court Judges (CACJ). Current data for this study includes the years 2020-2025, representing over 26,000 MHC participants. The sample (n= 2,964) is 61% Male and 39% Female, including White (56%), African American (40%), Hispanic (1.5%), Asian (0.4%), American Indian/Alaskan Native (0.4%), and other (1.9%) participants. Statistical analyses were conducted using R. Bivariate analyses show significant differences in risk/needs levels, incarceration length, and sanctions by disorder groups. Regression modelling will be used to assess whether disorder category predicts differences in program outcomes while controlling for relevant covariates. Specifically, I hypothesize that participants with externalizing disorders will experience higher rates of behavior-related supervision responses, and that participants with internalizing disorders will demonstrate lower levels of overt misconduct but may experience greater difficulty with consistent engagement.

Behavioral and Population-Level Responses of Colpoda steinii to Polystyrene Microplastic Exposure

Savanna Evans
Major: Microbiology
Faculty Mentor: Travis Ichikawa, New Materials Institute, College of Veterinary Medicine

Microplastic pollution is a growing global concern and has been investigated extensively in marine systems, yet its effects in terrestrial and freshwater environments need further elucidation. Colpoda steinii, a suspension-feeding bacterivorous ciliate common to both habitat types, may be particularly vulnerable to microplastic exposure due to the nature of its feeding behavior, which risks uptake of inert particles in lieu of nutritive prey. Polystyrene is a persistent synthetic polymer widely used in single-use packaging and consumer products and is a common source of microplastic contamination. This study investigates the effects of polystyrene microspheres on C. steinii population dynamics and movement behavior across a range of particle concentrations. Cultures were established with fluorescent GFP-expressing Escherichia coli as the sole bacterial food source alongside fluorescent polystyrene microspheres, with bacterial prey not reintroduced following initial depletion. Populations were monitored over a four-day period, and fluorescence microscopy was used to track particle–ciliate associations and confirm intravacuolar ingestion of polystyrene particles. Additionally, the effects of particle concentration on C. steinii movement behavior and its capacity to discriminate between bacterial prey and polystyrene particles were examined. These findings contribute to a growing body of work on microplastic impacts in non-marine protists and have broader implications for understanding trophic transfer of microplastics through microbial food webs.

Use of RFID Technology to Study the Response of Brown-headed Nuthatch (Sitta pusilla) to Urbanization

Grace Kelly
Majors: Fisheries and Wildlife Sciences, Psychology
Faculty Mentor: Jeffrey Hepinstall-Cymerman, Warnell School of Forestry and Natural Resources
Additional Co-Author: Maria Fernandez

The Brown-headed Nuthatch (Sitta pusilla, BHNU) is a pine-specialist songbird dependent on mature pine stands for foraging and nesting, making it vulnerable to habitat loss and fragmentation associated with urbanization. Despite this vulnerability, little is known about the species’ ability to adjust territorial behavior in urbanized landscapes, in part because traditional color-banding methods often provide limited data for this species. This study addresses two questions: (1) do BHNUs alter territory size, composition, or territoriality behavior along an urbanization gradient? and (2) can RFID (radio frequency identification) monitoring improve territory estimation compared to the traditional banding method? Five study sites with various levels of urbanization were identified, and two BHNU families were selected per site. Between two and five individuals per family were banded with unique color bands containing RFID microchips. Territory use will be quantified both through visual resighting of banded BHNUs to test the traditional method and through RFID detections at specialized RFID feeders, which will be moved weekly to randomized locations within a 100-meter radius of the central nesting point. Detections from both methods will be used to estimate territory size and evaluate method effectiveness. We predict that families in highly urbanized areas will occupy larger territories and exhibit greater territoriality than those in less disturbed habitats. We also expect RFID monitoring to yield more frequent detections and more precise territory estimates than visual resighting. Results will clarify behavioral responses of BHNUs to urbanization and assess RFID technology as a tool for studying territory use in difficult-to-monitor songbirds.

Longitudinal Associations Between Body Mass Index and Brain-Age Gap in Adolescence

Steven Kesser
Majors: Mathematics, Statistics
Faculty Mentor: Shana Adise, Nutritional Sciences, College of Family and Consumer Sciences
Additional Co-Authors: Jonatan Ottino-González, Miguel Angel Rivas Fernández

Childhood obesity has been associated with differences in the brain’s structure and function. Previously, we showed that these differences may be due to greater glial activity, presumably resulting from cell death due to neuroinflammation, but increased glial activity may also be normative developmental synaptic pruning (“i.e., maturation”). Because obesity has been linked to cognitive decline, understanding these associations may help identify developmental risk factors.

We investigated whether body mass index (BMI) was associated with brain-age gap (BAG), a measure of brain maturation, in 9-to-12-year-olds with healthy weight at baseline (n=2474; 48.3% female; 16.7% Hispanic) in the Adolescent Brain Cognitive Development Study. BAG is the difference between estimated brain age from structural magnetic resonance imaging (MRI) and chronological age. We hypothesized higher BMI would be associated with greater BAG. Using T1 MRI data, harmonized for scanner variation, we generated brain-age predictions from 3 models: AgeDiffuse, a Whitmore XGBoost model, and a preliminary Beck XGBoost model. Linear models tested BMI associations, adjusting for sex, puberty, parental education, and birthweight

Among youth who transitioned from healthy weight to overweight/obesity (n=273), higher BMI was associated with greater BAG in the Whitmore model (β=0.101-0.120, p<0.001), but not in AgeDiffuse or the Beck model. In the full healthy-baseline sample (n=2474) and among youth who remained healthy weight (n=2201), AgeDiffuse showed positive BMI-BAG patterns; Beck results were less consistent. These findings suggest model choice may shape conclusions about BMI and adolescent brain maturation.

Behavioral Phenotyping Chemogenetic Modulation of Olfactory Bulb Astrocytes in Mice

Rachel King
Major: Biomedical Physiology
Faculty Mentor: Shaolin Liu, Physiology and Pharmacology, College of Veterinary Medicine

Olfactory dysfunction or loss of smell is an early symptom of neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease, often preceding the cognitive decline and motor impairment that typically prompt diagnosis. The olfactory bulb (OB) performs the initial synaptic processing of odorant signals with its outermost layer, the glomerular layer (GL), serving as the first site of synaptic integrations. Astrocytes, the most abundant glial cells in the OB, have recently been shown to regulate synaptic activity, yet their role in olfactory processing remains poorly understood. We hypothesize that inhibition of astrocytic activity within the GL of the OB impairs odor detection and odor recognition memory. To selectively inhibit astrocytes, adeno-associated viral (AAV) vectors encoding the inhibitory Designer Receptors Exclusively Activated by Designer Drugs (DREADDs), hM4D(Gi), under the glial fibrillary acidic protein (GFAP) promoter were microinjected into the GL of six wild-type C57BL/6 female mice. Six control mice received the pAAV-GFAP104 construct. With mCherry protein serving as the fluorescent marker, immunohistochemistry and confocal imaging will be used to validate the expression of the viral constructs in astrocytes. Impact on olfactory behavioral outcomes will be assessed using the Buried Food Test to evaluate odor detection and the Three-Chamber Social Interaction Test to measure social odor recognition memory. By defining the functional role of glomerular astrocytes in olfactory processing, this study aims to provide mechanistic insight into the olfactory deficits observed in early stages of neurodegenerative diseases.

Analyzing the Structural Effects of Farnesylating CaaX-Motif Proteins

Spencer Krone
Major: Biochemistry & Molecular Biology
Faculty Mentor: Walter K. Schmidt, Biochemistry & Molecular Biology, Franklin College of Arts and Sciences
Additional Co-Author: Emily R. Hildebrant

J-domain proteins (JDPs) exist within almost every organism. Many JDPs are heat shock proteins (HSPs) that function as protein co-chaperones to facilitate folding, refolding, trafficking, remodeling, and disaggregation of other cellular proteins. Understanding the function of JDPs is extremely vital as their dysfunction is associated with neurodegenerative disorders (e.g., Alzheimer’s and Parkinson’s). This post-translational modification (PTM) involves covalent attachment of a farnesyl lipid to a protein’s CaaX-motif (CaaX proteins contain cysteine, two aliphatic residues, and a varied amino acid titled “X” at the C-terminus). The functions of farnesylated JDPs are regulated by this modification, but the molecular reason is unresolved. The goal of this study is to use Nuclear Magnetic Resonance (NMR) to investigate the hypothesis that farnesylation regulates the three-dimensional structure of farnesylated proteins. This study first required optimization of an E. coli system to produce various proteins (yeast Ydj1; human DnaJA1, DnaJA2 and Nap1) that receive farnesylation, which involved co-expression of target proteins and farnesyltransferase. This was followed by optimization of induction conditions to recombinantly produce Ydj1 with and without the modification. The production and percent farnesylation of Ydj1 associated with these conditions was determined by SDS-PAGE analysis of E. coli extracts followed by NIH ImageJ analysis of band intensities. Ydj1 has since been produced and purified using M9 minimal media containing 15N-NH4Cl, which is suitable for NMR studies. These labeled preparations of Ydj1 will enable structural comparisons of modified and unmodified Ydj1 to directly determine whether farnesylation alters the conformation of Ydj1 to potentially modulate its co-chaperone function.

The Role of Abraçada Cells as a Potential Niche for Stem Sell Function in Schmidtea mediterranea

Kathleen Miller
Major: Genetics
Faculty Mentor: Rachel Roberts-Galbraith, Cellular Biology, Franklin College of Arts and Sciences
Additional Co-Author: Skylar Settles

Regeneration describes the concept of incorporating new tissue after injury. Many organisms have varying levels of regenerative capabilities. Planarians are one such species, as they are capable of whole-body regeneration. Gap junction networks have been proposed as a possible mechanism for regeneration during injury and homeostasis. Innexins, the invertebrate ortholog of vertebrate connexins, are protein components which make up gap junctions that serve as communication networks between cells. We therefore hypothesize that cellular communication occurs between cells during stem cell maintenance and proliferation. It is unclear what cells provide this potential niche, which provides signals and support to the stem cells. We propose that in planarians abraçada cells serve as this potential niche cell. By utilizing RNA interference and in situ hybridization, we show that two innexins, inx8 and inx9, expressed in abraçada cells exhibit decreased stem cell expression and stem cell phenotypes. As these gap junction components are only a network by which signals cross, we also highlight the potential metabolites and metabolic pathways inhibited by the disruption of mRNA in vivo. We can then highlight the importance of four metabolite genes and a potential metabolic pathway necessary for stem cell maintenance and function. Understanding the cellular communication networks between stem cells and their potential niche would enable us to identify a pathway to manipulate maintenance both in vivo and in vitro. Furthermore, understanding how these mechanisms are conserved across organisms may enable us to translate this mechanism to other invertebrate and possibly vertebrate models.

Integrating Smartphone-based Photogrammetry and Reflected Satellite Signals on a UAV platform for Accurate Soil Moisture Sensing

Phong Nguyen
Majors: Electrical Engineering, Mathematics
Faculty Mentor: Mehmet Kurum, Electrical Engineering, College of Engineering

Measuring soil moisture content is an essential requirement in precision agriculture. Alongside the usage of soil probes, Global Navigation Satellite Systems Reflectometry (GNSS-R) has gained significant attention in retrieving geophysical properties of the Earth’s surface through recycling electromagnetic signals from satellite constellations. Built-in GNSS chips and antennas on smartphones have been proven as an easily accessible approach for sensing soil moisture, but additional environmental factors like vegetation levels can introduce errors. To address this, this project integrates Structure from Motion, a photogrammetry camera vision technique that estimates 3D structure from a set of 2D images using matching RGB points. Implementation of SfM through the smartphone’s camera can estimate vegetation levels without the need for external LiDAR or Multispectral sensors. A motor-powered rotating platform holding the smartphone is attached to the UAV, designed to reduce irregular radiation patterns of smartphone GNSS chips. This early stage of the project has been utilizing a camera drone to perform photogrammetry in place of the smartphone setup. Preliminary results have shown photogrammetry 3D reconstruction to be comparable to expensive LiDAR data. Next stages will focus on extensive data collection over agricultural fields. After the UAV with a smartphone attached is flown over a field, the collected GNSS-R data along with visual vegetation information serve as input for a RandomForest Machine Learning model to generate a heatmap estimating moisture levels. This project aims to demonstrate that incorporating photogrammetry on a drone-mounted platform effectively accounts for vegetation biomass, thereby improving the accuracy of soil moisture measurement in complex agricultural environments.

Evaluating the Function of Anti-NXT-2 Monoclonal Antibodies for the Treatment of Invasive Pulmonary Aspergillosis by Passive Transfer in a Murine Model

Casey Ellen Pac
Majors: Cellular Biology, Communication Studies
Faculty Mentor: Karen Norris, Infectious Diseases, College of Veterinary Medicine
Additional Co-Author: Taylor Chapman

Invasive fungal infections (IFIs) are a critical concern because there are currently no clinically approved vaccines, increasing anti-fungal drug resistance, and high mortality rates. The majority of IFIs are caused by Aspergillus, Candida, Pneumocystis, and Cryptococcus. Our lab has developed a ‘pan-fungal’ vaccine candidate, NXT-2, that is a 90-amino-acid pan-fungal consensus sequence made from the kexin-like sequences of Aspergillus, Candida, Pneumocystis, and Cryptococcus. Previously, our lab has demonstrated that NXT-2 vaccination elicits antibody-mediated protection in murine and non-human primate models of fungal infections. To further expand on antibody-mediated protection, our lab has developed a panel of anti-NXT-2 monoclonal antibodies (mAbs) generated from mice vaccinated with NXT-2. In this study, we selected the strongest candidates based on in vitro functional assays for evaluation in an immunosuppressed murine model of pulmonary aspergillosis. First, we conducted a pilot study with a single dose of antibody to assess the ability of the mAbs to prevent morbidity and mortality in pulmonary aspergillus infection in mice through passive transfer. Then we conducted a second study with a two-dose regimen. These findings provide a foundation for designing future studies evaluating NXT-2 monoclonal antibodies as passive immunotherapies for invasive fungal infections, particularly those caused by Aspergillus fumigatus.

RIVER-SENSE: A Low-Cost Intelligent Platform for Real-Time River Water-Quality Monitoring

Shivank Pandey
Major: Computer Science
Faculty Mentor: Avishek Dutta, Geology, Franklin College of Arts and Sciences

Rivers sustain ecosystems and human communities, yet the continuous water-quality data needed to protect them remains scarce. Conventional monitoring depends on infrequent grab sampling or professional-grade multi-parameter sondes that can cost tens of thousands of dollars, leaving transient pollution events undetected and dense sensor networks out of reach for resource-limited settings. RIVER-SENSE is a low-cost, field-deployable platform that continuously measures temperature, pH, turbidity, and total dissolved solids and transmits the readings to the cloud for real-time monitoring. This project develops the platform’s software and intelligence layer. A virtual-sensing framework estimates dissolved oxygen, an ecologically critical indicator, directly from the four measured parameters using a machine-learning model trained on regional water-quality records. This removes the need for a physical dissolved-oxygen probe, one of the most costly and maintenance-intensive components of commercial systems and one prone to fouling in long-term deployments. A cloud architecture supports the framework in production: field hardware streams readings to a managed database, and a hosted dashboard runs the prediction pipeline to deliver continuous dissolved-oxygen estimates and visualizations to stakeholders. Calibration and quality-control procedures validate the readings against reference instruments. By shifting computation to the cloud, the system keeps field power and processing demands low while delivering data quality that approaches professional-grade instrumentation at a fraction of the cost, making reliable real-time monitoring feasible where it has long been unaffordable. The platform supports United Nations Sustainable Development Goal 6.1, universal and equitable access to safe and affordable drinking water, by giving communities and agencies the timely information needed to detect contamination and manage water resources. RIVER-SENSE demonstrates how accessible software and cloud tools can extend advanced environmental monitoring to the settings that need it most.

Establishing a Neuronal Aging Model Induced by Amyloid Beta

Darsh Patel
Majors: Regenerative Bioscience, Economics, Applied Biotechnology
Faculty Mentor: Steven Stice, Regenerative Bioscience, College of Agricultural and Environmental Sciences
Additional Co-Author: Marzan Sarkar

Aging and Alzheimer’s disease are closely linked to the accumulation of amyloid beta 1-42 oligomers. However, studying human neuronal aging remains difficult because reliable, controllable models are scarce. This raises a question: can Aβ1-42 oligomers be used to induce neuronal aging in human stem cell-derived neurons in a predictable, dose-dependent way? We propose that Aβ1-42 oligomers represent key features of neuronal aging in human pluripotent stem cell-derived prefrontal cortical neurons, establishing a modifiable in vitro platform for studying neurodegeneration. To test this, we differentiated human pluripotent stem cells into prefrontal cortical neurons over a roughly forty-day protocol and confirmed neuronal identity through characteristic morphology. We prepared Aβ1-42 oligomers from HFIP-treated monomers and validated the toxic oligomeric species using A11 antibody slot blot analysis. Neurons were then treated across a range of Aβ1-42 concentrations, and we assessed neuronal morphology by bright-field imaging and cell viability by trypan blue exclusion. Our findings show that Aβ1-42 oligomers produce dose-dependent neurite retraction, cell loss, and reduced viability, indicating that the model reliably captures features of neuronal decline. Building on this foundation, we next aim to characterize the aging phenotype more deeply through mitochondrial and metabolic readouts and longevity-associated signaling, and to test whether neural stem cell-derived extracellular vesicles offer neuroprotection. Establishing a controllable human model of neuronal aging is significant because it provides a much-needed platform for investigating the mechanisms of age-related neurodegeneration and for screening anti-aging and neuroprotective therapeutics.

The Role of Heparan Sulfate in Extracellular Vesicle Binding and Uptake

Le Xuan Ai Phan
Major: Biochemistry & Molecular Biology
Faculty Mentor: Yao Yao, Animal & Dairy Science, College of Agricultural and Environmental Sciences
Additional Co-Authors: Rachel Hankin, Yaochao Zheng

Extracellular vesicles (EVs) are nanovesicles released by cells that can transport bioactive cargo across the blood–brain barrier, making them promising therapeutic tools for neurodegenerative diseases. A major challenge in EV-based therapy is improving brain targeting, which requires understanding the molecular mechanisms of EV uptake. Heparan sulfate (HS), a sulfated sugar on cell and EV surfaces, has been implicated in this internalization process. In this study, we investigated how cell-surface HS and specific sulfation patterns on both donor and recipient cells regulate EV binding and internalization. We hypothesize that distinct HS modifications modulate selective vesicle targeting.

Using CRISPR/Cas9-engineered knockout (KO) cell lines (EXT1, NDST1, HS2ST1, and HS6ST1/2 DKO), we evaluated the binding and uptake of NPC and HEK-derived EVs by various KO recipient cells, alongside the internalization of KO donor EVs by hepatoma (Hep3B) and neuroblastoma (SH-SY5Y) recipient cells. EVs were fluorescently labeled with CFSE to track and quantify binding and uptake via flow cytometry.

For recipient KOs, the total loss of HS and N-sulfation reduced NPC-EV uptake, while HEK-EV uptake decreased across all four lines. Conversely, HS2ST1 and HS6ST1/2 KO donor EVs significantly enhanced uptake in Hep3B cells, and HS6ST1/2 KO EVs promoted binding and uptake increases in neuroblastoma cells. These findings demonstrate that recipient-cell HS is vital for vesicle docking, whereas donor-surface 6-O-sulfation may act as an inhibitory barrier. Future research will analyze surface HS composition, expand recipient screening, and perform in vivo biodistribution tests.

Investigating Synthetic Genetic Interactions with DED1

Steven Schnell
Majors: Biochemistry & Molecular Biology, Spanish
Faculty Mentor: Timothy Bolger, Biochemistry & Molecular Biology, Franklin College of Arts and Sciences

In order to effectively adapt to environmental stressors, cells regulate their gene expression. Nearly every step in gene expression is facilitated by a highly conserved protein family called the DEAD-box proteins, but the way they control different subsets of mRNAs remains unclear. Ded1 is a DEAD-box protein with an important role in translational regulation. Under normal cell conditions, Ded1 upregulates translation by stimulating assembly of the 48S pre-initiation complex (PIC) and unwinding secondary structure in the 5’ untranslated regions of mRNAs. However, under stress conditions, Ded1 causes dissociation of its binding partner eIF4G1 and itself from the 48S PIC, resulting in degradation of both proteins. Interestingly, when the C-terminal region of Ded1 is removed (ded1-ΔCT), some amount of Ded1 and eIF4G1 is retained even under stress conditions and low levels of translation still occur. A previous genetic screen identified non-essential genes that genetically interact with ded1-ΔCT. Growth was compared between strains grown on nutrient rich media under normal cell conditions and strains grown on media containing rapamycin to mimic stress. Double mutants that grew better than mathematically expected under stress were labeled as enhancers of the ded1-ΔCT phenotype, while those that grew less than expected under stress conditions were designated as suppressors. Growth assays were conducted for four hits selected from the prior screening to confirm its accuracy. Initial results have both confirmed and contradicted the prior screening for different mutants. Future experiments via western blot will determine whether expected protein degradation occurs in the presence of rapamycin.

Redesigning Recycling: A Decentralized System for Converting Waste into Sustainable Composite Materials

Mackenzie Smith
Majors: International Business, Real Estate
Faculty Mentor: John Aliu, Engineering Education Transformations Institute, College of Engineering

How can recycling be redesigned to reduce environmental impacts while increasing the economic value of recovered materials? Current recycling systems rely on centralized facilities that require extensive transportation, energy-intensive sorting equipment, water-intensive cleaning processes, and multiple stages of material handling before recyclable materials can be reused. These processes increase operational costs, greenhouse gas emissions, and resource consumption while contributing to low recovery rates for many materials, particularly plastics. This research investigates the feasibility of a decentralized recycling system that processes waste at its point of generation. The proposed system seeks to eliminate unnecessary transportation and reduce dependence on water-intensive cleaning and centralized processing, creating a more efficient and sustainable recycling model. In addition to improving the recycling process, the project explores methods of increasing the value of recovered materials by combining plastic, paper, and aluminum into durable mixed-composite products, such as low-cost construction bricks and other building materials. By transforming traditionally low-value waste into functional products, the system aims to support a more circular economy while reducing landfill disposal and resource consumption. The expected outcome is a scalable proof-of-concept that demonstrates how localized recycling and value-added manufacturing can improve material recovery, lower operating costs, reduce carbon emissions, and create affordable, sustainable products with potential applications for universities, communities, and offices as well as multiple other industries.

Engineering tRNA Anticodon Stem Structure to Improve Stop Codon Selectivity: The Role of Stop Codon Position

Alina Soifer
Majors: Regenerative Bioscience, Chemistry
Faculty Mentor: Natalie Krahn, Biochemistry & Molecular Biology, Franklin College of Arts and Sciences

Proteins are normally synthesized using 20 standard amino acids, but genetic code expansion (GCE) enables the incorporation of non-canonical amino acids (ncAAs), such as pyrrolysine (Pyl), to create proteins with novel properties. These engineered proteins have applications in biotechnology and therapeutics. A major challenge in GCE is achieving selective suppression of a target stop codon while minimizing suppression at non-target codons. Improving this selectivity is important for increasing the reliability of engineered translation systems.

This project investigated how stop codon position influences suppression selectivity and efficiency in GCE systems. Previous studies suggested that UAG codons were more selective for Pyl incorporation than UGA codons. However, those experiments differed in stop codon position, making it unclear if this led to the observed results. This raised the question of whether stop codon position contributed to the observed selectivity differences rather than stop codon identity itself.

To study codon identity and position, green fluorescent protein (GFP) reporters containing either UAG or UGA stop codons at positions 2 or 151 were compared using pyrrolysine tRNAs carrying either CUA (UAG) or UCA (UGA) anticodons. Across four independent assays, UGA codons exhibited greater selectivity than UAG codons regardless of stop-codon position. These findings indicated. The contrast between these results and previous studies suggests that promoter choice may significantly influence measurements of stop codon selectivity.

Understanding the determinants of stop codon selectivity is imperative for more precise genetic code expansion technologies, research, and biomedical applications.

Does PAC Ideological Composition Predict Competitiveness in U.S. House Races?

Hayden Trebon
Majors: Finance, Political Science
Faculty Mentor: Jamie Carson, Political Science, School of Public and International Affairs

Political action committees (PACs) are commonly assumed to sort into competitive U.S. House races along ideological lines, with “ideological” PACs (e.g., Club for Growth, EMILY’s List) concentrating resources to swing outcomes while “access” PACs hedge across both candidates to preserve influence regardless of winner. Despite the prominence of this narrative in campaign finance scholarship and political commentary, it has rarely been tested directly against race-level PAC giving patterns. This study asks whether the ideological composition of PAC contributions predicts electoral margin in competitive House races once district partisanship and incumbency are held constant. Using FEC bulk contribution data, DIME v4.0 PAC ideology scores, Cook Partisan Voting Index (PVI), and MIT Election Lab returns, I constructed a race-level panel of U.S. House races decided by 10 points or fewer from 2016 to 2024 (n = 320) and calculated two race-level measures: a Hedge Intensity Index (HII), capturing the share of PAC dollars given to both major-party candidates within a cycle, and a PAC Composition Index (PCI), capturing the ideological tilt of each side’s PAC coalition under both contribution-share and DIME CFscore-based specifications. Regressing general-election margin on HII and PCI, controlling for absolute PVI, incumbency, open-seat status, and total PAC spending, I find that neither HII (p = 0.20) nor PCI (p = 0.72) significantly predicts margin, with the full model explaining only a modest share of variance (R² = 0.19); this null result is robust to alternative competitiveness cutoffs, exclusion of party committees and leadership PACs, and the inclusion of cycle fixed effects. These findings suggest that once district lean and incumbency are accounted for, PAC ideological composition does not meaningfully explain how close a House race turns out to be, challenging a common assumption in campaign finance discourse and pointing toward outcome variables beyond margin, such as turnout or post-election legislative behavior, as more promising tests of whether PAC ideology has electoral consequences.

Editorial Values and Economic Realities in UK-Based Independent Music Journalism

Stella Turner
Majors: Journalism, English
Faculty Mentor: Kyser Lough, Journalism, Grady College of Journalism and Mass Communication

This study uses in-depth interviews with editors of independent online music publications in the United Kingdom to explore the sustainability pressures of music journalism and how these outlets see their role in shaping cultural visibility. This involves a discussion of culture in how editors choose what to publish and whom to amplify, under the broader sustainability pressures of how these decisions are constrained or shaped by revenue models, labor conditions, and survival strategies.

Integrating Sustainable Technologies into Airport Architecture: Energy Efficiency and Aesthetic Design Across Climate Zones

Veronika Viazovaia
Major: Mechanical Engineering
Faculty Mentor: Matthew Ryan Riggs, College of Environment and Design

With the rise of development in aviation infrastructure, new ecological problems have arisen. Because of this, humanity needs to consider sustainable strategies to mitigate the negative environmental impact of airport facilities. Understanding climate zones and weather conditions allows for the implementation of effective types of technologies in the exact region. In many existing studies, sustainability is explored as something separate from the architecture. With this in mind, how can sustainability be integrated architecturally into airports across different climate zones? How can energy consumption be reduced while enhancing aesthetic quality?

This research examined six airports across three climate zones. This study analyzed the reduction of ecological impact through 35 charts. These charts documented water saving options, solar panel placement and usage, and LED facilities within the three climate regions. This data was collected through open sources and Google Earth GIS, which served as the basis for further calculations. The comparative analysis identified relationships between climate, airport scale, and passenger density, revealing the most effective technology for each context and the broader patterns of their implementation.

After analyzing the effect of technological integration on airport aesthetics, the study documented the intersections between sustainability and architecture, emphasizing specific technology based on the climate zone.

The result generates useful findings for both aviation and architectural fields, showing how new technologies may upgrade the existing airports and which may be beneficial when considering the economic implications associated with creating a new airport.

Carbonatites of Mt. Vulture: Decoding an Italian Magmatic Anomaly

Trista Warner
Major: Geology
Faculty Mentor: Mattia Pistone, Geology, Franklin College of Arts and Sciences
Additional Co-Authors: Claudia Romano, Alessandro Vona, Danilo Di Genova, Fabrizio Di Fiore, Jacopo Taddeucci

Carbonatites are among the rarest and most unusual magma types on Earth. Rich in carbonate minerals (essentially liquid limestone) they flow as quickly as water when hot, but cool and crystallize so rapidly, they almost always freeze and stall out underground. The ancient Mt. Vulture volcano in southern Italy is a major geological anomaly because these rare magmas actually managed to breach the surface and erupt.

To understand how carbonatitic magmas are erupted, we collected representative rock samples at Mt. Vulture and we carried out laboratory investigation simulating heating and foaming of an alkali-rich carbonatite rock. Thermogravimetric Analysis (TGA) revealed a massive, rapid weight loss of 10% as the sample approached 1100°C. This critical weight loss documents the explosive release of trapped volcanic gases (CO2, which was part of the carbonate miner composition which is CaCO3), a process that dramatically strips the magma of its volatiles. To study the physical impact of this transformation, the degassed material was melted to synthesize a volcanic glass whose composition evolved into an andesite (common intermediate magma composition in the Earth’s crust) after volatile loss. Subsequent Differential Scanning Calorimetry (DSC) successfully pinned down the precise temperature window where this material locks up, revealing a distinct glass transition temperature interval between 662°C and 700°C followed by immediate crystallization at the nanoscale.

By capturing the exact weight loss and thermal transition points of this material, we map how gas loss alters magma movement. Following, Scanning Electron Microscopy (SEM) will be used to allow us to visually inspect these experimental textures, creating a benchmark to reconstruct the extreme conditions required for gas-charged carbonatite magmas to erupt globally.

Chemical Kinetics of Tetrahydropyran

Raiah Wright
Major: Biochemical Engineering
Faculty Mentor: Brandon Rotavera, Chemistry and Engineering, Franklin College of Arts and Sciences, College of Engineering
Additional Co-Author: Troy Smith

The transportation sector remains heavily reliant on energy-dense liquid fuels, with petroleum-derived hydrocarbons and oxygenated biofuels providing 99.5% of its energy demands. The U.S. Energy Information Agency (EIA) projects this trend to continue, with 95% of energy needs in 2050 coming from hydrocarbons and biofuels. To address this, low-temperature combustion (LTC) engines have emerged as a promising candidate for next-generation transportation. These engines operate at temperatures and fuel-air ratios that minimize emissions and accommodate a variety of alternative fuels, including second-generation biofuels, which are produced by deconstructing biomass cell walls. However, a major barrier to the widespread adoption of LTC technology is the need for more accurate chemical kinetics models under broad ranges of temperature and pressure. Such models enable a thorough understanding of the complex chemical pathways involved in biofuel combustion

To address knowledge gaps in understanding chemical reactivity, this research focuses on the theoretical characterization of reaction steps for tetrahydropyran, a typical second-generation biofuel, to provide quantum chemistry data for combustion modeling. In total, seventy-four stationary points were computed, gathering their corresponding single-point energies and vibrational frequencies resulting in 262 calculations. These structures were then organized into AutoStorage, a flexible database for quantum chemical calculations, to establish a centralized repository. Future work will focus on computing the transitionary states involved in tetrahydropyran combustion, allowing us to generate a master-equation input file. Ultimately, this data enables us to integrate and expand existing kinetic models, replace incomplete mechanisms, and advance the development of low-temperature combustion fuels and engine technologies.

Investigating Heparan Sulfate Biosynthesis Through Enzyme Turnover

Duojia “Jenny” Wu
Majors: Biochemistry & Molecular Biology
Faculty Mentor: Ryan Weiss, Biochemistry & Molecular Biology, Franklin College of Arts and Sciences

Heparan sulfate (HS) is a highly sulfated polysaccharide expressed on the surface and extracellular matrix of all mammalian cells, where it regulates cell signaling, development, coagulation, and tissue organization. The structural diversity of HS chains is determined by the coordinated activity of HS biosynthetic enzymes in the Golgi apparatus, and the dysregulation of this process has been implicated in neurodegenerative tauopathies, skeletal disorders, and cancer. Despite our knowledge of the function and activity of enzymes responsible for HS polymerization and modification, the mechanisms dictating HS polysaccharide assembly and sulfation patterning remain poorly understood. We hypothesized that the turnover of individual Golgi-localized enzymes directly influences HS sulfation patterning and downstream biological functions. To investigate this, we generated CRISPR-mediated HiBiT reporter cell lines in HeLa cells for five key HS biosynthetic enzymes, HS6ST1, HS3ST1, HS2ST1, GLCE, and NDST1, thus enabling endogenous detection and quantification of protein levels using highly sensitive luminescence-based assays. Our initial studies focused on H6ST1-HiBiT cells demonstrated that HS6ST1 protein turns over within 24 hours, similarly to the HS polymerase enzyme exostosin-2 (EXT2). Based on these findings, future experiments will characterize the turnover kinetics and degradation pathway of all five HS enzymes to determine how the HS biosynthetic machinery is regulated during biosynthesis. Overall, these findings have broad implications for understanding how cells control HS structure and may inform therapeutic strategies for diseases related to HS dysregulation.

How Does Salmonella Grow on Diced Vidalia Onions at Different Temperatures?

Che Yang
Major: Food Science
Faculty Mentor: Abhinav Mishra, Food Science & Technology, College of Agricultural & Environmental Sciences
Additional Co-Authors: Harsimran Kaur Kapoor, Juan Gao, Aishani Tewari

Salmonella is a major pathogen present in food products. While Salmonella outbreaks are the most common in foods of animal origin, onions have been a significant source of Salmonella in recent years. From 2018 to 2023, the CDC reported 2,590 illnesses, 501 hospitalizations, and 1 death from Salmonella linked to nine outbreaks from onions. Among different onion varieties, Georgia’s famous Vidalia generates between 120 and 200 million dollars per year. Compared to standard varieties with around 5% sugar, Vidalia onions contain at least 12% sugar, which may facilitate the growth of microbes such as Salmonella. In particular, diced onion presents a higher risk because of its higher moisture content and reduced protection form the skin. This project seeks to establish a predictive mathematical model describing Salmonella growth on diced Vidalia onions, specifically exploring the effect of different temperatures. Diced fresh Vidalia onions are inoculated with a 5-strain Salmonella cocktail containing strains from past produce-linked outbreaks, the onions are then incubated at different temperature, and Salmonella counts are enumerated at different time points using serial dilutions and spread plating on xylose lysine deoxycholate (XLD) agar. The growth data is fitted to a Baranyi primary model at each temperature to estimate the growth parameters at each temperature. In the next step of the project, a modified Ratkowsky secondary model will be used to examine the effect of temperature on the growth parameters; finally, a dynamic model will be established and validated with a programmable incubator following real-world supply chain temperature patterns.