Skip to content

Our latest publication

The Spring 2024 Issue

Volume 22, Issue 2
ISSN: 1931-9983 (online) | ISSN: 1547-9641 (print)

Selected Abstracts from the Inaugural Undergraduate Lightning Talks 2024

University of Rochester, Rochester, NY
April 10th, 2024
Losing Diversity Through Hybridization? Distinct Populations of the Flame-Rumped Tanager Vanishing by Homogenization

Author: Elizabeth Croyle ‘24

Mentor: J. Albert Uy and Maria Castaño

Incipient species boundaries can change over time due to gene flow and geographic isolation. These distinctions can blur, collapsing back into one population, or be reinforced, finishing the speciation process. Using this framework, I am exploring the consequences of gene flow and isolation between two subspecies of a South American songbird. Throughout its range in Colombia, Ramphocelus flammigerus flammigerus and R. f. icteronotus are separated by the Andes but come into contact (sympatry) in at least three independent transects, where they produce a range of hybrids in areas of low-elevation mountain passes. To investigate the demographic history of the two subspecies and the effects of gene flow on divergence, I am using two coalescent, computational methods: 1) popsizeABC and 2) fastsimcoal2.  First, I am using popsizeABC to estimate changes in effective population sizes up to 1 million years in the past, and investigating any historical decreases or increases. Second, I am testing different demographic history scenarios using fastsimcoal modeling to determine the most likely scenario that matches our observed transect 2 genomic data. My results show that the demographic model of divergence in sympatry (subspecies living in the same area) with increasing gene flow between subspecies is the most likely scenario for transect 2. Therefore, I predict that the subspecies will continue to become more similar, and eventually collapse back into a single population. Populations share a similar demographic history, with a prominent bottleneck around 6500 years ago, followed by rapid expansions.  The timing of the bottleneck coincides with dramatic decreases in population sizes found in pre-Colombian human populations, most likely caused by unstable climatic conditions in the region.  Next, I will run fastsimcoal on the remaining two transects and assess if sympatric homogenization is supported and if the demographic parameters (time of divergence, magnitude and directionality of geneflow) are consistent.

The Democratization of Art through Technology: A Viennese Case Study

Author: Robert Marcinauskis ‘24

This project examines the democratization of the art market through digital instruments that unlock new revenue sources and investors. Technologies, such as NFTs and art shares, have brought a closer financial analysis of art as an asset. Comparatively, the blue-chip market has been subject to the historical arbitrary valuation of physical art pieces. Democratization occurs through the institutional validation of these instruments; such is the case of the Belvedere Museum. Pieces of art that have not been on the market for centuries are now accessible assets. In addition, the careful tracking of painting prices allows for stable, anticipated returns. The research conducted is through case studies focusing on the modern technologies explored. The primary case focuses on the partnership of artèQ and the Belvedere Museum in Vienna, Austria. This partnership created NFT’s tiles of Gustav Klimt’s The Kiss, which have unlocked new revenues for the museum and more investment opportunities for the art piece. This project challenges the uncertainty of digital instruments with the practical exposure of new investors to an asset class that had limited itself to an elite class. The cultural center of Vienna provides a backdrop to this revolution of Art, with its monarchical roots limiting artistic exposure of poorer classes. With the validation received by the Belvedere, digital instruments become an important asset class that expands exposure to the returns of the art world.

Bridging Accessibility Gaps: Developing The First COVID-19 Health Chatbot Tailored to the Deaf

Author: Akber Ahmed ‘24

Mentor: Jose Perez Ramos

In order to bridge the gap in creating more accessible technologies, my research project pioneers the development of the first COVID-19 health information chatbot tailored specifically to the deaf community. This utilizes a comprehensive approach that integrates UX Research methods and Human-Centered Design principles. Through different phases of group testing of the prototype, the chatbot is carefully developed to meet the needs of deaf users, ensuring accessibility and usability. Guided by a mission to blend technological innovations, user experience research, and translational science seamlessly, we aim to tackle health disparities and meet community needs through innovative Information Communication Technology (ICTs) solutions. By integrating expertise in UX research, translational science, and ICT innovation, our goal is to close technological access gaps for underserved communities, driving more impactful technological advancements.

Serglycin Regulates Azurophilic Granule Formation in Neutrophils

Author: Wimeth Dissanayake

Mentor: Julia von Blume

Neutrophil granules are membrane-bound compartments that release their contents upon a stimulus in a process that is called regulated secretion. Neutrophils contain three prevalent granule types, characterized by different protein compositions and capacity to fuse with the cell membrane: primary (azurophilic), secondary (specific), and tertiary (gelatinase) granules. The proper secretion and formation of Neutrophil granule proteins protect the body against infections, inflammation, and malignancies. However, the underlying molecular mechanisms of granule biogenesis in Neutrophils are poorly understood, limiting the treatment of neutropenic disorders. Granule proteins are synthesized, folded, and quality-checked before their translocation to the Golgi apparatus, with subsequent sorting to granules occurring in the trans-Golgi Network (TGN). The process by which granule-destined proteins are sorted and subsequently packaged into nascent secretory granules has remained a scientific enigma. Azurophilic granules (primary granules) are the first granules formed during Neutrophil differentiation and contain serine proteases, neutrophil elastase (Elane), and myeloperoxidases (Mpo). Previous studies suggested that Serglycin, Cd63, and Rab27a are required for the trafficking and function of azurophilic granules. Our unpublished work confirmed the increased production of Serglycin during the initial stages of neutrophil differentiation, corresponding with the initial expression of Elane, indicating that Serglycin may play a role in Azurophilic granule formation. To test if Serglycin is a granule-forming protein, we ectopically expressed serglycin in U2OS cells naturally deficient of secretory granules. Our data showed that Sergycin induces Elane, Rab27a, and Cd63 expression. Furthermore, immunofluorescence microscopy of Granulocyte-monocyte progenitors (GMPs), imaged at different time points of differentiation, showed increased expression and colocalization of Serglycin, Elane, Cd63, and Mpo during azurophilic granule formation. Our findings support the theory that Serglycin regulates azurophilic granule biogenesis by promoting the upregulation of cargo proteins (Elane) and trafficking machinery, including Cd63 and Rab27a. With this work, we have provided a first mechanistic insight into the biogenesis of azurophilic granules, setting the base for future investigations.

Measuring Data Access Latency in Large CPU Caches

Author: Shaotong Sun ‘24

Mentor: Chen Ding

This talk describes a new, multi-locality benchmark program for testing memory access latency and using it to study recent AMD machines equipped with 3D vertical cache (V-Cache) that can be over 1 GiB in total size on a single node. The latency study shows that these large caches differ from traditional LLCs in two aspects: the V-Cache is partitioned rather than shared, and the cache replacement policy is more similar to random than it is to LRU.

SiC_lightning_talk – Shaotong Sun

The Effect of Modifications to DNA Structure on the Architecture of Condensed Bacterial DNA with a Nucleoid-associated Protein

Author: Kevin Zheng ‘24

Mentor: Anne Meyer

Escherichia coli increases the abundance of DNA-binding protein from starved cells (Dps) in response to various stressors bacteria encounter when exposed to a variety of environments. One of the primary functions of Dps is to bind and condense DNA. In in vitro systems with different buffer conditions, different structures of Dps:DNA condensate have been observed. Our lab has termed these condensates spongiform, globular, and liquid-like. My research focuses on the contribution of different features of DNA to the morphologies of Dps: DNA condensates. DNA behavior and interactions differ over different length scales, supercoiling, and methylation states. Modifying these features of the DNA could alter the accessibility of the negatively charged DNA backbone and positively charged tail of Dps or alter the typical DNA structure, causing changes to the later formed condensates. To test the effect of varying DNA lengths, we incubated PCR products of different lengths (linear DNA fragments) with Dps to form condensates. The DNA in these condensates was mixed with fluorescently tagged DNA hairpins and Dps was mixed with Dps variants chemically labeled with Alexa fluor 647, enabling both to be visualized simultaneously. We used fluorescence microscopy to analyze the resulting morphologies. Our findings show that while changing the length of the DNA at the concentration we use does affect the size and number of condensates, the overall structure didn’t vary. This suggests that shorter DNA may condense more effectively, while longer DNA creates larger structures. Further studies regarding the effects of DNA supercoiling and methylation states will provide a better understanding of the major factors that play into condensate formation.

Towards Deciphering the Mechanism of Microglia-Driven Remodeling in Brain Extracellular Space Using SFA-FRAP Imaging

Author: Showmick Ranjan Paul ‘25

Mentor: Dr. Edward Brown

Learning and memory require dynamic movement of different diffusible molecules and structural changes of dendritic spines, crucial for synaptic plasticity and the formation of new neural connections. However, being a physical barrier, the extracellular matrix (ECM) of the brain can affect this movement. We hypothesize microglia as a master manipulator of the ECM due to the remarkable motility of microglial processes, and the ECM-degrading enzymes that they produce. We utilize acute brain slices from mice and apply a quantitative optical method, Fluorescence Recovery After Photobleaching with Spatial Fourier Analysis (SFA-FRAP), which we have been developing and optimizing specifically for this investigation, to measure diffusive transport within ECM across various microglial motility states, including activated, deactivated, and absent conditions. Our objective is to assess whether microglia actively modulate diffusive transport and, if so, to elucidate the precise mechanism through which they accomplish such modulation. Results from this study are anticipated to shed light on the mechanisms underlying microglia-ECM interactions and provide insights into potential therapeutic strategies for neurological disorders associated with aberrant synaptic plasticity.

Legal Accessibility for Domestic Violence Survivors

Author: Carly Zubrzycki ‘26

Mentor: Rachel O’Donnell

Willow Domestic Violence Center offers a variety of services to domestic violence survivors such as counseling, shelter, and legal advocacy. For the purposes of my project, I focused on increasing capacity for their legal advocacy program as there is limited capacity to keep community members informed of legal processes and resources by meeting with them one on one as the program typically entails. Likewise, there is a lack of resources in the local Rochester area that accept pro-bono or sliding scale domestic violence cases, particularly ones without a stringent income cutoff or enough variety to prevent conflict of interest. Thus, there is a need to create resource guides for common legal processes and establish pro-bono and sliding scale connections. The objective of the project was to establish 1-3 legal guides and develop at least one pro-bono or sliding scale connection with a local lawyer. In addition, the aim was to create a transition guide as well as a resource document so these resources can exist beyond my work. In order to create the legal guides, I completed training days at Willow to observe what types of legal guides are needed, researched legal processes, and consulted my supervisors for review. To facilitate pro-bono and sliding scale connections, I conducted a literature review for best practices of contact and then I reached out to lawyers via email and phone to inquire about logistics and the extent of the partnership. For the resources document, I researched local area resources and use cases. Ultimately, these capacity building measures serve to increase legal accessibility by providing resources and services for those in need.

Lightning Talk (Carly Zubrzycki) – Carly Zubrzycki

The Effect of Language Framing on Climate Change Perception

Author: Hanna Felber ‘24

Mentor: Karen Berger

Climate change is an increasingly pressing issue, but, despite its urgency, there is difficulty in garnering public support to address it. Many people still don’t care, or worse, oppose the concept entirely. Because of this, it is crucial to understand how to communicate climate change information in a way that encourages people to reduce their climate footprints and support large scale efforts to decrease emissions. Previous research has revealed that the use of framing and metaphorical language can be extremely effective in shifting opinion, especially when used in media. This study applies these findings to climate change language, discovering just how influential framing is when discussing this topic. The participants, all US citizens between the ages of 18-35, first completed a pre-survey regarding their belief of climate change and willingness to act on it. They then read a paragraph on climate change that either used metaphorical language, non metaphorical language, accountability-based language, or non-accountability-based language. Lastly, they completed a post-survey with the same questions as the pre-survey. To measure the impact of framing, participants’ pre- and post-survey responses were compared between the four conditions. Changes in response between these two surveys indicates an effect of language framing on climate change belief. The findings reveal that when discussing the harms of climate change and individual responsibility, non-metaphorical non-accountability language is more effective to use than metaphorical non-accountability language. Additionally, individuals were more likely to believe that they made effective actions to address climate change when presented with metaphorical non-accountability language. While further research is needed, these findings support the notion that framing does impact climate change perception, and therefore must be utilized to garner support.

Synergistic Effects of an HDAC Inhibitor, a Therapeutic Cancer Vaccine, and an Immunocytokine in an HPV Tumor Model

Author: Sree Chatterjee ‘26

Mentor: Jeffery Schlom

There are over 630,000 new cases of HPV-related cancers worldwide each year. Despite the increasing prevalence of prophylactic vaccines as therapeutic treatment options, their impact on cancer incidence will be minimal for several decades. The only currently approved immunotherapy is anti-PD-1/PD- L1 in cervical and head and neck squamous cell carcinoma (HNSCC), whose response rate is only 10-20%. This project aimed to determine if three anti-cancer agents, PDS0101 (a prophylactic cancer vaccine), NHS-IL12 (histone targeting antibody), and Entinostat (HDAC inhibitor) used in combination have potential as a therapeutic treatment option for HPV-related cancer patients. We specifically looked at effects of the triple combination therapy on chemokine & cytokine levels within the tumor microenvironment & periphery, as well as links between chemokine/cytokine levels and anti- tumor response & immune cell populations in the tumor. We found that the triple combination therapy yielded the strongest anti-tumor effect, and that NHS-IL12 was a driving factor for increasing chemokine/cytokine levels. We also found consistency between MIG levels in tumor supernatant and CD8 levels in tumor, as well as KC levels and G-MDSC levels. Finally, high levels of immune stimulating cytokines were present within groups treated with some combination of agents (double & triple). However, the chemokine/cytokine data alone did not explain the level of anti-tumor response within triple therapy, so continued study is needed to determine synergistic effects of each agent.

Tagmentation-based genomic mapping of Tol2 transgene insertion sites in zebrafish

Author: Paige Schneider ‘26

Mentor: Patrick Murphy

Transgenesis is a pivotal technique for unraveling gene regulatory networks and developmental processes. While the Tol2 transposon-based transgenesis system stands as the predominant method for generating transgenic lines in zebrafish, exact insertion sites for the majority of those generated lines remain unidentified. The lack of insertion site information raises concerns about potential local influences on transgene expression, off-target effects at protein-coding genes, and complications arising from copy number variations. To address these issues, we introduce Tol2Mapping, a novel method utilizing Tn5 transposase-mediated tagmentation for streamlined identification of genomic insertion sites for zebrafish Tol2 transgenic lines. Tol2Mapping is not only straightforward to perform but also allows for the identification of insertion sites from heterozygous transgenic lines and the detection of multiple insertion events. We also provide a detailed protocol for Tol2Mapping library generation, step-by-step sequencing analysis, and a user-friendly Shiny app, making this method accessible to researchers without programming expertise. Altogether, Tol2Mapping emerges as a valuable tool to enhance the precision and utility of zebrafish transgenesis studies, addressing critical challenges when using Tol2 transgenic lines.

Selected Abstracts from the Annual Undergraduate Speakers Symposium 2024

University of Rochester, Rochester, NY
April 5th, 2024
Natural Sciences
Identifying Early Biomarkers in Sclerotic Vulvar Disease

Author: Ashley Updike ‘24

Mentor: Megan Falsetta

Lichen sclerosus (LS) is a chronic inflammatory and scarring disease commonly found in the anogenital region. LS affects male and female anatomy, but 85% of cases are identified on the vulva (external part of the female genitalia). The current estimated prevalence is as great as 3% in women. At least 5 million US women live with LS. The most common symptoms in are pain, vulvar pruritus (itching), dysuria (difficulty or pain while urinating), and dyspareunia (pain during intercourse). Over time, this leads to severe changes in the vulvar architecture that can only be altered by physically removing parts of the vulva, which may be necessary to restore voiding function (ability to urinate) and relieve other symptoms. The exact prevalence of LS is unknown and underreported. It is imperative that we find disease biomarkers that could be used to identify LS before these irreversible and potentially life-threatening architectural changes occur. Our hypothesis is there are genes that are highly expressed in the regions adjacent to the visibly scarred regions in LS patients that could serve as unique biomarkers for early disease diagnosis. To look at the temporal effects of gene expression in perilesional areas, a time course experiment was performed and the preliminary data is consistent with my hypothesis that there are early disease markers. The genes we have identified could serve as biomarkers for disease to prevent permanent architectural changes and ultimately cancer. Steroids are effective for most patients, but they are ineffective if delivered after permanent architectural changes occur. These findings could be directly translated to clinical applications to improve quality of life for many patients.

Coarse-Grained Strategy for Modeling the Self-Assembly of Janus Nucleobases

Author: Derek Chien ‘25

Mentor: Ignacio Franco

Janus nucleobases are two-faced hydrogen-bonding heterocycles that undergo hierarchical and programmable self-assembly to form biomimetic supramolecular polymer networks—such as nanotube structures composed of stacked hexameric rosettes—with extensive biomedical applications. However, the mechanism and driving forces that dictate the non-covalent self-assembly of these molecules remain poorly understood, preventing the rational design of Janus nucleobases to guide the self-assembly into specific outcomes. Here we develop a computational strategy to investigate the thermodynamic driving forces directing the hierarchical self-assembly of hybrid guanine-cytosine (G∧C) Janus nucleobases. We favor a coarse-grained strategy, as opposed to fully atomistic modeling, in order to computationally access the experimentally relevant system size and time scales that lead to self-assembly. In the model, the G∧C nucleotide is represented as an patchy isosceles trapezoidal prism in which three types of facets allow for selectively directional interactions: two complementary Watson-Crick pairing edges and the aromatic planar face responsible for stacking. MP2 first-principle electronic structure computations are carried out on G∧C module pairs in relative positions and orientations that exemplify hydrogen bonding or stacking; the resultant energies are used to parameterize the patchy particle model and accurately replicate the pairwise interactions between the coarse-grained facets. In turn, the self-assembly process is modeled through Monte Carlo simulations using the coarse-grained model. The strategy is able to reproduce the self-organization of G∧C modules into rosette nanotubes and thereby informs about the importance of hydrogen-bonding and stacking interactions on modulating the hierarchical self-assembly.

Neural Mechanisms Contributing to Auditory Processing Disorders in Developmental Disabilities

Author: Sarah Mehta ‘25

Mentor: Emily Knight

Children with ADHD (attention-deficit/hyperactivity disorder) and ASD (autism spectrum disorder) struggle in understanding speech in noisy settings, causing difficulty in academic and social scenarios. Understanding the neural mechanisms causing these auditory processing disorders may inform future interventions in this area. Using simultaneous 3D virtual reality (VR) and electroencephalogram (EEG) integration, we have developed a novel experimental paradigm allowing the naturalistic study of how two proposed neural mechanisms (audiovisual integration and selective visuospatial attention) may influence auditory processing in children with ASD, ADHD, and typical development (TD) ages 7-12 years. The stimuli consists of three virtual characters: a center, target character and two flanking distractor characters presented in VR. The participant is told to press a button every time the center person says a target word, ignoring any words spoken by the distractor characters. We manipulate audiovisual integration by including randomly intermixed trials that are visual-only (character lip movements only), audio-only (character speech sounds only) and audiovisual. Spatial attention demands are manipulated with non-interference trials (where only the center character is speaking) and interference trials (where all three characters are simultaneously speaking). We measure steady state visual evoked potentials (SSVEPs) through high density EEG by tagging the target and distracting speakers with visual frequency oscillations (23 hz and 15 hz respectively). If the strength of corresponding SSVEP bands correlate to 23 Hz and/or 15 Hz, we are able to determine where visual attention was located. Preliminary behavioral results in children with TD (n=7), ADHD (n=4), ASD (n=1) and ASD+ADHD (n=1) reveal highest accuracy across groups in the non-interference condition. Relative to TD children, children with ASD or ADHD may show smaller performance gains in the audiovisual compared to audio-only condition. Preliminary EEG data suggests there may be a trend emerging toward potential group differences in allocation of visuospatial attention. Children with TD tend to show robust SSVEP primarily in the 23 Hz (target) band while those with ASD or ADHD tend to demonstrate SSVEP in both the 15 Hz (distractor) and 23 Hz (target) frequency band. Taken together, if these trends persist with continued data collection, this could imply that audiovisual integration and visuospatial deficits may contribute to auditory processing disorders in children with ASD and/or ADHD.

How parasites manipulate neural gene expression of developing hosts

Author: Natasha Vacca ‘25

Mentor: Floria Mora-Kepfer Uy

Across animal systems, hosts and their parasites display a strong arms-race where parasites are incredibly successful pathogens that exploit the host via less understood mechanisms. To study these mechanisms, our lab recently found gene expression patterns in the brain of adult social wasp hosts, Polistes fuscatus, that reflect parasitic manipulation via their insect parasite Xenos peckii. Parasites infect female workers destined to be short-lived and expand host life span 9-13 times longer while creating a neural gene expression profile that is similar to long-lived queens. However, it has been unclear if these parasites begin manipulation as their hosts develop, or instead they only manipulate at the adult level. I developed a project to determine when parasitic manipulation commenced in host neuronal pathways using neuronal RNA extraction from infected worker larvae, uninfected worker larvae, and uninfected future queen larvae. My first results show differential gene expression between infected and uninfected larvae that mirrors the altered expression of future queens compared to workers during development (Fig. 1). In early stages of development, the parasite manipulates key genes associated with immune regulation, aging, and metabolism to alter the host’s overall neural gene expression to match long lived future queens in its last instar. Additionally, the patterns of development from these hosts show an asynchronous development within hosts until the last parasite larval stage also supporting this novel finding in gene expression modulation. These findings shed light on how this emerging aging model system develops and subsequently reprograms hosts via parasitic neuromodulation over time.

Engineering & Math
Detecting AI-Generated Essays: Unveiling Bias Against AI In Residency Application Evaluations

Author: Loralai Crawford ‘24 / Take 5 ‘25

Mentor: Nicole Wilson

Introduction: Chat Generative Pre-Trained Transformer (ChatGPT) is one of many currently available large language models (LLMs) that can generate responses to user input. Applicants are increasingly using these tools to build their residency applications. However, based on internal discussions, program directors and faculty may consider use of artificial intelligence (AI)-generated responses a form of cheating. As LLMs improve, the ability of faculty to detect AI-generated text is questionable. Our purpose was to determine whether faculty of a general surgery residency program can detect AI vs. human-written responses to a text prompt and determine how assumed essay source affects reviewers’ ratings.  We hypothesized that faculty would not be able to reliably differentiate AI vs. human-written responses and that bias against presumed AI-generated essays would exist.

Methods: 10 faculty reviewers were recruited from the general surgery residency program. A standard text prompt, “Please tell us in 1-2 paragraphs why you are considering [our institution] for General Surgery residency” was used to prompt 10 essays, 5 from current trainees and 5 from GPT3.5 or GPT4.  AI responses were selected as is and included in review packets after de-identification. Blinded reviewers rated essays (10-point Likert scale) for: desire to interview the applicant, relevance to general surgery residency, and overall impression; and indicated whether each essay was AI- or human-generated. Scores and identification error rates were compared between the two AI and human written essay groups using Mann-Whitney (median±interquartile range) and logistic regression (AUC±std error).  Essays were redivided based on how they were identified by reviewers and total points awarded were compared between groups using Mann-Whitney (median±interquartile range) and 2-way analysis of variance (ANOVA).

Results: There was no difference between groups for % total points awarded (ChatGPT 66.0±13.5%, Human 70.0±23.0%, p=0.508, Figure 1). Higher overall impression scores were associated with reviewer identification of an essay as human-generated (AUC 0.82±0.04, p<0.001). Except for one essay, all essays were identified incorrectly by at least 2 reviewers. There was no difference between groups for identification error rates (ChatGPT 40.0±35.0%, Human 20.0±30.0%, p=0.175, Figure B). Essays thought to be AI-generated (57.0±13.8%) scored lower in % total points awarded than those perceived to be human-written (71.5±9.3%, p = 0.014)

Conclusions: Applicants use AI tools in all academic settings, but controversy exists regarding whether this should constitute a form of academic dishonesty. Our results demonstrate that while humans and AI tools generated essays of similar quality, bias exists against presumed AI-generated essays. Interestingly, faculty were not able to reliably differentiate human from AI-generated essays, suggesting the bias may be misdirected. AI-tools are becoming ubiquitous and their use is not easily detected. Therefore, faculty must expect these tools to begin to play larger roles in medical education.

Atomic-scale Simulation of Nanowires Under Bending Deformation

Author: Sicheng(JD) Qian ‘25

Mentor: Niaz Abdolrahim

Body-centered-cubic (bcc) metallic materials, especially Molybdenum (Mo), are used in a significant number of applications due to their excellent mechanical properties. Components made by nanomaterials for various uses usually encounter both simple uniaxial stress and complex stress conditions. Current studies have focused on material properties, such as phase transformation and twin boundary, in Mo metallic nanowires (NWs) under uniaxial tensile and compressive loading. However, only a limited amount of research has focused on bending deformation in BCC material due to the technical difficulty of the experiment. This research utilizes LAMMPS to compute the MD simulation of bending deformation on the BCC Mo NWs. A complete phase transformation path from bcc1 to face-centered-cubic (fcc) to bcc2 has been observed in the <1 0 0> and <1 1 0> Mo NWs. Tetra-twin boundary (TTB) forms in the structure along {1 1 2} slip system under bi-directional bending. In contrast, only {1 1 0} slip system has been activated under mono-directional bending. As a result, nucleation of TTB enables second elasticity and large uniform plastic deformation in the structure. This study reveals a new type of twining structure under non-uniaxial loading and largely fulfills the understanding of Mo and BCC material mechanical behaviors.

Enhanced Reliability Predictions: An AI-based Framework for Performing Failure Modes, Effects, and Criticality Analysis in Industrial Environments

Author: Stephanie Wang ‘25

Mentor: Frank Zou

Reliability engineering grapples with the imperative task of predicting and comprehending product failures across an array of sectors. Although traditional approaches to ensure product safety in reliability engineering like Failure Mode Effects and Analysis (FMEA) and Failure Mode, Effects, and Criticality Analysis (FMECA) are successful, their inherent manual intensity and reliance on expert insights introduce limitations. As today’s systems surge in complexity, a more holistic approach becomes paramount. In response, we introduce an AI-driven risk assessment tool that guides the user to a host of failure modes and their effects for each component contained in a bigger system. Through a user-friendly graphical interface and a robust statistical modeling backend, the AI-driven tool streamlines the risk assessment process by prompting users to input a system’s name and subsequently generating an extensive array of failure modes and associated effects for each constituent component, including Weibull, Rayleigh, and Bathtub distribution curves. By automating this aspect of FMEA/FMECA, the AI-based solution seeks to not only enhance reliability analyses but also optimize development timelines, improve resource allocation, and provide valuable educational avenues for junior across sectors, including chemical, automotive, aerospace, and beyond.

Challenging the Stereotype of the Witch: Medea in 15th Century Burgundy

Author: Kirsten Bell ‘24

Mentor: Laura Smoller

One specific trend that occurred in fourteenth and fifteenth century Europe was a shift in attitude towards witchcraft and magic, in that it hardened and eventually resulted in the multitude of witch trials of the sixteenth and seventeenth centuries. This shift also included the formation of the diabolical witch stereotype, where ideas about dangerous witches were no longer isolated to clerical scholarly circles but spread to the public. One way that this seems to be shown is through the literature published by some members of the Church discussing how dangerous witches are to society. In regards to literature not published by the Church, I found that stories featuring Medea, one of the most famous mythological ancient enchantresses, can indicate how these societies viewed magic and witchcraft. In the ancient world, Medea was a fictional princess of Colchis who used magic to assist the hero Jason in obtaining the Golden Fleece. In this myth, Jason later betrays her by marrying another woman. When conducting research on the medieval literature where Medea appears, one book stood out, titled L’Histoire de Jason. This was written in 1460 by Raoul Lefèvre and commissioned by Duke Philippe le Bon of Burgundy. At the same time that this book was written, an incident occurred that can be considered one of the first institutionalized witch trials. These trials occurred in the city of Arras in the region of Burgundy between the autumn of 1459 and 1460. They were led by Dominican friars who attempted to equate witchcraft with apostasy, which is when one renounces Christianity and either cannot repent for their sins or only do so with great difficulty. In this thesis, I argue that Medea’s specific portrayal in Lefèvre’s L’ Histoire de Jason reflects the attitudes around witchcraft and magic in Burgundy in the late 15th century, in that she is a redeemable magic user but harmful magic is still a problem. It indicates echoes of ancient and earlier medieval conceptions of magic, where magic is a tool that can be used for both good and evil. Medea is not inherently evil for using magic because she is always driven by her immense love for Jason, and reconciles with Jason at the end of the story after she becomes a penitent. Since Dominican friars attempted to equate witchcraft with apostasy through the Arras incidents, the fact that Medea is able to repent for her sins shows how 15th century Burgundians did not completely buy into this new conception of demonic and non-repentable witchcraft.

Sensory Stimuli: An Examination of Light, Sound, and Movement

Author: Glory Linebach ‘24

Mentor: Anne Wilcox

Sound and light are utilized to create atmosphere, mood, and tone by artists. When generating a work of art, the creator takes into account what colors or sounds they should incorporate to evoke the emotion of their work. In dance specifically, choreographers use light and sound to create an atmosphere around the choreography. That said, how does the choreographer know what specific colors and sounds will go with specific movements to create the mood they are looking for? Current research within the field of color theory has shown strong connections linking light to specific emotional responses. In other words, specific colors represent emotions (MSEd, K). Similarly, in the field of brain-and-cognitive science, different pitches have also been correlated to cause emotional responses; however, the range of emotions that connect to a sound frequency span a more basic spectrum (happy to angry) compared to the emotions related to a color of light (Vos S). In recent research, movement analysts have found that specific movement vocabulary and qualities of movement are connected with the expression of emotions: specific movements correspond with the emotion a person feels in that moment (Tracy). My research proposes that we can connect the emotion felt with the color of light, the pitch of sound, and the subject’s movement qualities. Multiple workshops were conducted to collect data and investigate this correlation. A dance was then created to highlight the results of the relationship between light, sound, emotion, and movement.

From Belief to Retreat: Prophetic Failure in the Fifth Crusade

Author: Clarice Speight McKee ‘24

Mentor: Laura Smoller

Paul Veyne famously asked the question “Did the Greeks believe in their myths?” I want to ask a version of this question about medieval Europe. Chronicles and letters from medieval Europe abound in mythical tales and prophetic texts, often conveniently aligning with the author’s own polemical goals. Were these just cynical manipulations, a literary trope, or, to echo Veyne, did medieval Europeans believe their prophecies? I argue that, during the Fifth Crusade (1217-1221), they did. Sent to reconquer Egypt from the Ayyubid Caliphate, the Fifth Crusade ended in miserable failure. Arriving in Egypt in 1218, the Crusaders were a ready audience for tales about an eastern Christian king named Prester John, whose kingdom was wealthy and pious beyond the European imagination. They received prophecies, both Christian and Islamic in origin, that appeared to promise that Prester John was only weeks away, ready to save the Crusaders from the Ayyubid Caliphate’s forces and liberate the holy city of Jerusalem from the “infidels.” Simultaneously, a text known as the Relatio de Davide began circulating among the Crusaders, describing the central Asian conquests of a descendant of Prester John, King David, which sounded quite like the real conquests of the Mongol empire. This only reinforced their belief that Prester John was coming to rescue them. I argue that their belief in these tales explains one of the puzzles of the Fifth Crusade, the Crusaders’ seemingly odd choice to advance into Egypt in the summer of 1221, just before the annual Nile flood would cut off any means of escape. Why retreat when reinforcements were on their way? Of course, instead of receiving aid from the mythical Prester John or King David, crusading armies were trapped by the flooded river and soundly defeated. Chronicles produced after the Fifth Crusade repudiated their eager belief in the tales about Prester John, now arguing that the prophecies may have been wrong and that Prester John may not even have been Christian after all. That chroniclers felt the need to ‘correct’ the Prester John myth rather than ignore it shows the continuing importance of the myth. The continuing search to understand who Prester John was shows that yes, medieval Europeans really did believe in their myths.

Unveiling the Pixels: A Critical Analysis of Black Representation in Action Video Games

Author: Kendal Jordan ‘25

Mentor: Kristana Textor

The depiction of black and African identities in video games has been neglected due to societal perceptions of the ‘main character’ – white, male, muscular, and religious. This content analysis explores the representation of the Black experience in AAA action video games, addressing the evolution of stereotypes and their influence on player perceptions. Drawing on historical context, from Jim Crow Era stereotypes to contemporary portrayals, the study reveals a transition from overt to insidious forms of racism within gaming narratives. Previous literature findings indicate a systematic overrepresentation of white, male protagonists, with Black characters often relegated to stereotypical roles, props for abuse, or sidelined entirely. This study aims to highlight the perpetuation of learned biases and stereotypes through gaming experiences (e.g., affecting adolescents’ perceptions of race and gender). Through a sample of 10 prominent titles, characters were selected and assessed based on various criteria, including education, gender, and positive role portrayal. To get a better grasp of the overall opinions of the game, 4 comments from each game were collected from Steam – a digital storefront. 48 participants completed a survey that collected their general views on the character’s design. We found that certain characters (e.g., Sheva, Everett, and Marcus) had the most similar survey responses to the Steam comments of their respective games. The main takeaway of this research is to explain the importance of diverse representation both in character portrayal and within the gaming industry, through advocating for increased inclusion and empowerment of marginalized voices. This is a preliminary analysis so a bigger sample size and more varied games are needed to delve deeper into the complexities of intersecting representations.

Social Sciences
Testing the Effects of Artificial Intelligence Disclaimers in Political Advertising on Viewer Perceptions

Author: Garrett Briggs ‘24

Recent polls demonstrate that American voters are distrusting of artificial intelligence. Simultaneously, lawmakers have started to regulate artificial intelligence (AI) in political advertising, with states passing legislation requiring the disclosure of AI assistance through a visual disclaimer. Surprisingly, no research explores whether attributions to artificial intelligence in campaign material affects viewers’ reactions. My research investigates this gap through a video experiment testing the effects of partisan border wall ads with and without AI disclaimers. I find that advertisements are more trusted in the presence of AI disclaimers, with the differences being significant across four outcome measures. Additionally, I find that the results are heavily influenced by partisanship, with Democrats being substantially more likely to reward AI disclaimers than Republicans. These results provide valuable information to policymakers and candidates surrounding the implications of AI-assistance disclaimers in political advertising.

Mediated Memory and Propaganda in an Analysis of the United States Embargo on Cuba

Author: Dariel Guerra ‘25

Mentor: Molly Ball

This study examines how United States propaganda, Cuban state-run media, and mediated memories have shaped the understanding around the United States embargo on Cuba. The argument I make is that the United States embargo on Cuba is the culmination of Cold War misleading narratives and unproductive politics disseminated through various media. This examination of the embargo relies heavily on print sources (newspapers, advertisements, posters) and digital media sources (blogs and newspapers). I analyze existing historical, political, and legal literature through an investigative lens of mediated memory and propaganda. How have hegemonic reconstructions of Cuba in memory manifested into hard-line policies against the Cuban government? How does propaganda play into the reconfigured, imagined Cuba in the minds of both Americans and Cuban-Americans? Utilizing the methodological framework from Aggressive and Passive Propaganda: Cuba and the United States by Eugene H. Pons I rhetorically investigate trends, views, and sentiments emerging from the media covering the embargo. In recent years, discourse surrounding Cuba has significantly diminished despite the ongoing continuation of the embargo and the strengthening of sanctions in 2017. The two nations continue to antagonize one another through media, disseminating narratives of half truths through newspapers, online independent news outlets, and blogs. My research concludes that the United States embargo on Cuba is a culmination of Cold War media warfare and aggressive propaganda. As a result of this media warfare, communities in the U.S. have distorted perceptions of Cuba, therefore leading to the continuation of the United States embargo. These modern representations of Cuba reinforce false narratives that ultimately fall on the backs of the Cuban people. Future work will explore the historical significance of the decades in Cuba preceding the Revolution, specifically examining case studies of United States involvement under the Platt Amendment (1901-1934). These case studies underlie greater trends that have provided subtext for the Revolution and the embargo, as well as ensuing media wars between Cuba and the United States.

Evaluating Interventions to Increase Cancer Screening Follow-Up Rates in High-Risk Emergency Department Patients

Author: Trisha Mondal ‘27

Mentor: Beau Abar

Background and Objectives: Emergency departments (ED) play a vital role in addressing community health needs, serving as a unique gateway to preventive services like cervical, colorectal, and lung cancer screening. Interventions involving research associates/volunteers can be used to promote screening in the ED, though differing patient populations across screenings may necessitate differential efforts at intervention and retention. Our study compares follow-up rates to three preceding clinical trial pilots focused on cancer screening for cervical (females ages 21-65), colorectal (adults ages 45-75), and lung cancer (adults ages 50-80 with a 20 pack-year smoking history who currently smoke or quit smoking within the last 15 years). We aim to assess the extent to which retention efforts might require adaptation across cancer screening populations.

Methods: Separate clinical trial pilot studies were conducted for the three cancer screening types, evaluating the effectiveness of a text-based behavioral intervention versus a verbal referral with written materials through the ED. In the current analysis examining study retention, participants’ outcomes were classified as successfully reached or lost to follow-up based on response to a phone call to self-report their screening data 150 days after their initial ED visit. Up to 5 call attempts were made for each participant. Data were compared across studies using chi-square tests for independence.

Results: In the colorectal and cervical cancer pilots, 61% (69/114) and 57% (54/95) of participants, respectively, were successfully reached at follow-up. Despite using the same methods, only 44% (88/198) of participants in the lung cancer pilot were reached via phone (p< 0.001). No significant difference in reachability was observed between colorectal and cervical cancer screening. These disparities cannot be attributed to sex, given nearly identical response rates in colorectal (men and women) and cervical (women) cancers, or COVID (timelines of the colorectal and lung cancer studies largely overlapped).

Conclusion: Patients requiring lung cancer screening exhibited the lowest likelihood of answering follow-up calls, posing a challenge in evaluating interventions in this specific population. Our findings highlight the need for more targeted efforts in retaining lung cancer screening patients to enhance the efficacy of preventive health care measures.


  • Santonastaso, B., Spector, H., Cahill, C., Akwaa, F., Ifthikharuddin, J. J., Gim, G., Refaai, M. (2024). Severe bleeding due to an acquired FXIII inhibitor in an otherwise healthy patient. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Ou, X., McGrath, K., Kingsley, P., Palis, J. (2024). Investigating the Interaction between Macrophages and Erythroid Cells in Bone Marrow: Implications for Erythropoiesis and Anemia. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Tamm, N.L., Miller, R. (2024). NMDA Receptors in Stroke: Pathways and Potential Treatments. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Bischoff, M.A., Wang, X., Song, E., Wilson, G., Brody, M., Iyer, S., Caine, K., Clark, M. (2024). Martha Chase at the University of Rochester: The Woman in STEM Who Was Forgotten. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Frank, R., Jarvis, M. (2024). Frederick Douglass: Man of Pictures. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Heberle, D., Heuer, C. (2024). Voyeur & Violator: The Obscene Narrative in Early Modern Italy. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Kolers, A., Pedersen, J. (2024). Aging Well: Politics of Aging, Sexuality, and Womanhood in The Golden Girls. University of Rochester, Journal of Undergraduate Research, 22(2).
  • Various Authors. (2024). Lightning Talks/Speakers Symposium Title. University of Rochester, Journal of Undergraduate Research, 22(2).

Journal Staff


Michael Christof

Wimeth Dissanayake

Managing Editors

Adya Mohapatra

Jason Lin

Marvin Wu

Rishika Chikoti

Content Editors

Sudarshan Ramanan

Anna Eisenberg

Audrey Chung

Yunji Kim

Trisha Mondal

Dina Garber

Allison Kuck

ZhuZhu Cai

Associate Editors

Manita Opoku

Nadia Pentolino

Isabel Doshi

Pauline Gordula

Maddie Greene

Owen Milewski

Amelia Ostrow

Julia White

Layout Head

Catherine Lan

Cover Designer

Phoebe Shin

Layout Editors

Amii Matamoros Delgado

Anagha Nair

Hanna Wang

Vismitha Prashanth

Churan Wang

Kayla Morgan

Zeina Omar El Madani


Office of Undergraduate Research

Emily Rendek, Director

Ann Robinson

Maureen Gabalski

Professor Sina Ghaemmaghami

The College

Jeffery Runner, Dean

Dustin Hannum, PhD

Professor Review Board

Linda Callahan

Bruce Blaine

Katherine Grzesik

Laura Smoller

Theodore Brown

Ning Wang

The Journal of Undergraduate Research | Creative Commons Attribution 4.0 BY International License
Return to the top of the page