48440 research outputs found
Sort by
Bacterial Genome Engineering for Lead (PbII) Aptamer Expression
No full textLead (II) is a pervasive toxin with serious neurological and developmental effects. There are currently very limited treatments that can reverse or prevent lead toxicity, many of which have significant side effects. Aptamers are short, single-stranded DNA or RNA sequences that can bind to specific substrates, including lead ions. Our lab has shown that aptamers have potential as prophylactic or therapeutic interventions to lead toxicity. The goal of this project was to test the lead-sequestering capability of an engineered E. coli strain that produces the lead-binding aptamer, Pb-7S. Herein I show that these engineered cells can produce the RNA Pb-7S aptamer, though do not accumulate Pb-7S upon induction. Regardless, the strain was tested for its ability to sequester lead. The results suggest some increase in lead sequestration capacity in the Pb-7S strain, however we cannot yet conclude that the lead sequestration effect is a direct result of Pb-7S expression. Further, I demonstrate that a bacterial biosensor based on the lead-regulated transcriptional repressor PbrR is capable of accurately measuring lead in aqueous samples in the parts per billion range. The work represents the foundation for future research to develop aptamer technologies in probiotic bacterial strains to sequester lead from the human gut
Interactivity and Worldbuilding: Intersections through Audio Dramas, Live Action Role Plays, Tabletop Role Playing Games, and Narrative Archival Research Puzzles
No full textThis paper is the culmination of work from August 2024 - May 2025. Throughout this time, our team researched, created, and worked towards the production of four interactive medias, exploring the emergence of interactivity and worldbuilding. These projects include an Audio Drama, Live Action Role-play, Narrative Archival Research Puzzle, and Tabletop Roleplaying Game, each centered around part of Khalyn, our overall world. Through this project, we sought to answer the question of how interactivity and worldbuilding intersect within selective interactive media
2024: CS: MQP: Walls: The Pwnable Claw Machine v2
No full textThis project explores the integration of AI-powered large language models (LLMs) into cybersecurity education to address gaps in student understanding and engagement, particularly in complex, technical domains. Traditional educational tools often fall short in delivering personalized, real-time feedback, which can hinder learning in asynchronous or large classroom settings. By leveraging LLMs' capabilities to provide context-aware, conversational guidance, we developed CEDRA—a tutor bot designed to support students through cybersecurity challenges with tailored feedback and expert-like reasoning. In addition, we designed a series of hands-on hardware security challenges to give students practical experience in embedded systems security, complementing software-focused instruction and preparing them for real-world cybersecurity threats. Together, these innovations aim to make cybersecurity education more interactive, accessible, and effective
Enhancing Electric Vehicle Efficiency through Advanced Power Electronics and Regenerative Braking Optimization
No full textThis project focuses on the design, thermal analysis, and real-world testing of a DC-DC boost converter intended to support energy efficiency in electric vehicle (EV) power systems. The converter was modeled and simulated using MATLAB, LTspice, and SolidWorks to predict electrical behavior and thermal performance. Key components such as the inductor, MOSFET, diode, and capacitor—were selected for their electrical ratings and thermal properties. Two physical prototypes were developed: one on a breadboard using discrete components, and one using a commercial PCB-based module. Thermal testing revealed the diode as the hottest component, prompting further analysis of heat dissipation across the system. Efficiency and temperature readings were collected from both prototypes to compare theoretical performance with real-world outcomes. The results validate the feasibility of compact, efficient voltage regulation, offering insight into improved thermal management for power electronics in electric vehicles
Self-sensing Construction Materials
No full textStructural health monitoring is used to evaluate the safety of a building. The leading way to determine the health of structure is to estimate the actual stress and strain in a structural member. Self-sensing concrete experiences a reduction in electrical resistance when exposed to compressive forces, therefore stresses and strains can be predicted. The prohibitive factor in conductive concrete is that conductive fillers can have a big impact on the production cost of the mix design and can weaken the maximum compressive strength of the concrete. This research focuses on developing a cost-effective solution to replacing carbon black with biochar which demonstrated similar conductive behavior and increased compressive strength compared to carbon black
Identifying novel genes required for neuronal cilia morphogenesis
No full textTranscription factors (TFs) are essential proteins for regulating gene expression and play crucial roles in brain function and disease. The Regulatory Factor binding to the X-box (RFX) family of TFs is best known for regulating transcription of core cilia genes. RFX TFs are highly expressed in the brain with RFX1, RFX3 and RFX7 showing particularly high expression in regions such as the prefrontal cortex, hippocampus, and cerebellum. Multiple RFX TFs have been implicated in neurodevelopmental and neuropsychiatric disorders, however, their brain-specific targets have not been systematically examined. This study aims to identify new brain-specific target genes of RFX TFs (1-7) using bioinformatics approaches. To this end, correlation analyses (Spearman, Pearson, and Maximal Information Coefficient) were performed using the RNA-seq expression data from 966 human brain samples acquired from The Human Protein Atlas to assess relationships between expression of full-length isoforms of RFX TFs (1-7) and 205,541 protein-coding transcripts in the human transcriptome. The analysis identified potential RFX-regulated genes, including ITSN1, NCAM1 and NCKAP1 linked to autism spectrum disorder (ASD). These results may contribute to a better understanding of RFX-mediated transcriptional regulation in the brain and its implications for neurological disorders and ciliopathies
Neural Imaging Classification and Feature Importance
No full textThis study compares multiple machine learning approaches for analyzing neural imaging data from the ASH neuron of \textit{Caenorhabditis elegans}. We applied random forests, multilayer perceptrons, and denoising autoencoders to classify neural responses to different chemical stimuli and identify the most informative temporal features of these responses. All three methods successfully classified stimuli at rates above chance level. Independent analyses from both the random forest feature importance metrics and denoising autoencoder masking experiments converged on the same finding: the immediate post-stimulus period (6.1-6.9 seconds) contains disproportionate predictive power for stimulus classification. This finding challenges traditional neuroscience approaches that focus primarily on neural response intensity and duration measurements. Our results demonstrate how machine learning methods can reveal previously unrecognized patterns in neural activity, with implications for both experimental design in neuroscience and explainable machine learning for biological data analysis
Standard Error Estimates for the Survivor Average Treatment Effect
No full textThis project presents a novel method for estimating treatment effects in trials with high attrition or dropout rates. Specifically, the study builds on a principal stratification approach to estimate the average treatment effect for students who would take the post-test regardless of treatment. Previous work developed causal estimators for principal scores, and this report extends that by introducing a method to calculate the standard errors of these scores and effects. The study applies this approach to an educational experiment evaluating two online learning programs among middle school students during the COVID-19 pandemic
Rare Earth, Rare Solution: Structural Insights into Synthetic Lanthanide-Sequestering Aptamers
No full textLanthanides are essential to modern technologies yet pose significant environmental challenges due to unsustainable mining practices. As constant demand grows for rare earth elements, biomolecular tools are needed for their selective sequestration. This study investigates the structural and functional properties of a 28-nucleotide long DNA aptamer, LnA_28, designed to bind lanthanide ions. Using a combined approach of computational modeling and experimental validation, this research explores how sequence modifications at key nucleotide residues (C15 and G19) impact aptamer conformation and metal binding. Molecular dynamics simulations showed changes in structural stability upon mutation of key residues in aptamer constructs. Circular dichroism spectroscopy confirmed metal-induced conformational shifts under varying pH and ligand concentrations. Among the variants tested, the G19 mutant exhibited the greatest effect on structural stability, suggesting its critical role in lanthanide ion coordination. These findings offer insights into the structure-function relationship governing aptamer-metal interactions and suggest that aptamer-based tools may contribute to sustainable lanthanide recovery
CS MQP - NVIDIA Chip Simulation
No full textThis project focuses on enhancing NVIDIA’s Tegra chip simulator (SCSIM) by integrating the existing simulator checkpoint-restore capabilities with Docker containerization. Working with NVIDIA’s Tegra Systems Software team, we enabled snapshot-restore functionality for newer System on Chip (SoC) models and integrated this feature within Docker containers. The implementation allows engineers to save simulator states and restore them later, significantly reducing debugging time when tests fail. Through rigorous testing, we validated system persistence across snapshots, including TCP connections, file integrity, and kernel module states. Most integration tests passed successfully, but we also identified areas for future improvement, particularly around managing multiple simultaneous simulations and handling volatile files. This work streamlines the chip development workflow by addressing common setup issues and reducing test execution time