ABCDEFGHIJKLMNOPQRSTUV
1
First NameLast NameAdvisorProject TitleAbstractTimeZoom Breakout Room
2
ShreyaAbhijitJahanikiaDietary patterns during COVID-19: Analyzing changes in consumption habits in adults before, during, and after COVID-19 lockdown"The COVID-19 pandemic and the resulting quarantine have caused drastic changes in an individual’s lifestyle. Dietary lifestyle, relating to an individual’s relationship with both food and exercise, is a major part of this change. We hypothesize that the pandemic caused participants to change their consumption patterns drastically, and that these changes have remained in place even after the lockdown period ended. Our questionnaire, which borrows ideas from four standardized eating assessments (SCOFF, EAT-26, CET, and EDE-Q) will aim to test this in adult participants. After completing data collection, we will conduct data analysis using R and various statistical softwares. The results of this study will provide a better understanding of one of the several changes that the pandemic has brought upon us."7:408
3
ShreyaAbhijitJahanikiaPsyGPT.ai: Developing an accessible, online source of comprehensive mental health services.Currently, receiving a diagnosis and/or getting therapy for a mental health disorder is not extremely accessible and generally requires multiple doctor visits. We hope to make getting fast and effective therapy easier and more accessible for those who are unable to or don’t feel comfortable visiting a therapist/doctor. Notably, veterans are at a significantly high risk for certain disorders, but mental health resources for this population are still lacking and relatively slow despite the vital importance of swift treatment. Our product aims to assist with the effects of and eventual diagnosis of depression and bipolar disorder. Though the two seem similar, their medical treatments and prognoses are extremely different. We hope to provide a user-friendly, interactive location for anyone to easily access available online mental health resources, such as the suicide and crisis hotline and various online tests.To achieve our goal of creating a successful AI powered-website, we need to create a proper design and develop the API, or, Application Programming Interface. We utilized Figma, a cloud-based design and prototyping tool to produce a user-friendly, approachable design for our chatbot. It is crucial to also design the main frame using HTML and CSS. With the help of Bootstrap, we hope to understand and construct a user-friendly website that can accommodate ChatGPT’s API keys. Once we create our website, we can import ChatGPT and train/test our data, developing a finished product.7:306STud
4
DikshyaAdhikariJahanikiaPsyGPT.ai: Developing an accessible, online source of comprehensive mental health services.Currently, receiving a diagnosis and/or getting therapy for a mental health disorder is not extremely accessible and generally requires multiple doctor visits. We hope to make getting fast and effective therapy easier and more accessible for those who are unable to or don’t feel comfortable visiting a therapist/doctor. Notably, veterans are at a significantly high risk for certain disorders, but mental health resources for this population are still lacking and relatively slow despite the vital importance of swift treatment. Our product aims to assist with the effects of and eventual diagnosis of depression and bipolar disorder. Though the two seem similar, their medical treatments and prognoses are extremely different. We hope to provide a user-friendly, interactive location for anyone to easily access available online mental health resources, such as the suicide and crisis hotline and various online tests.To achieve our goal of creating a successful AI powered-website, we need to create a proper design and develop the API, or, Application Programming Interface. We utilized Figma, a cloud-based design and prototyping tool to produce a user-friendly, approachable design for our chatbot. It is crucial to also design the main frame using HTML and CSS. With the help of Bootstrap, we hope to understand and construct a user-friendly website that can accommodate ChatGPT’s API keys. Once we create our website, we can import ChatGPT and train/test our data, developing a finished product.7:306
5
AnanyaAggrawalMcMahanA Comparison of Supervised Learning and Deep Reinforcement Learning for Autonomous DrivingAutonomous vehicles are a concept and field of vast interest, such as in the field of robotics engineering and in industries such as automobiles. MLAV1's objective is to identify the best machine learning method (supervised learning v.s. deep reinforcement learning) to train autonomous vehicles and thus achieve greater car safety. To compare the machine learning models’ performance, we built a miniature car (using raspberry pi and various sensors) modeled after an autonomous vehicle as well as a maze and obstacles. The machine learning method that leads to the car completing the maze the fastest, will be identified as the more efficient and accurate.7:104
6
AmizhthiniArivazhaganBrahInvestigating the interaction between clozapine and norclozapine with HLA proteins in the pathogenesis of clozapine-induced agranulocytosisClozapine-induced agranulocytosis (CLIA), is a rare but consequential reaction associated with the clozapine medication for resistant schizophrenia. Various classes of the HLA gene, which helps the immune system detect foreign pathogens, have been found to be associated with an increased risk of CLIA. However, the specific underlying mechanism remains unclear. We postulated that clozapine, or its metabolite norclozapine, may be interacting with HLA proteins to produce the intracellular aberrations that ultimately lead to CLIA. Our research revealed that specific HLA-B protein complexes increase the risk of CLIA by showing a stronger affinity for clozapine and norclozapine, offering new insights into treatment for resistant schizophrenia.7:301
7
AmiArivuAdamsAlgae powered batteryOur group is using Synochocytis algae to power a battery, seeing how long the home grown algae will power the battery compared to past trials7:001
8
DhanushArunmahasenanJahanikiaPersonalityGPT: Closing Communication Gaps in Texting using Artificial IntelligenceIn the realm of AI-driven communication, PersonalityGPT is a groundbreaking solution that leveraging the influential Big Five personality traits to bridge communication gaps across various personality types. This innovative model integrates with users' emotional states, aided by sentient AI components, enhancing empathy and understanding. By seamlessly adapting its responses to different personality types, the model provides more authentic and meaningful conversations. Furthermore, the integration of sentient AI components amplifies its empathetic capabilities, enabling it to evaluate and respond to subtle emotional states, thereby fostering a deeper understanding of users' mental health, a field that is becoming ever more important.7:205
9
UrviAvadhaniJahanikiaInvestigating the Impact of Meditation on Neural Processes: Insights from EEG and Sleep AnalysisThe popularity in the research of the impact of meditation on brain activity has spiked over the past few years. Even so, analysis on the effects of mediation through electroencephalography (EEG) has been done for decades but its impacts are still uncertain. This is due to how various meditation practices affect brain activity differently, shown evident in our dataset. The dataset included data from four blocks of meditation: two thinking blocks, one breathing block, and one tradition specific meditation block. The first meditation set is a breath count meditation which may contain sleep data. The first thinking task will work as the control data to compare with the breathing task. After sorting through 50 subjects, and running it through our software titled EEGLab, the data had to be pre-processed to make accurate conclusions. With continuing data analysis through EEGLab, and identifying sleep pattern in our data using various filtering and computing methods, we hope to find the direct effects of meditation and how it may benefit cognition, perception, and emotional processing.7:507
10
AnirudhAyyadevaraJahanikiaPersonalityGPT: Closing Communication Gaps in Texting using Artificial IntelligenceIn the realm of AI-driven communication, PersonalityGPT is a groundbreaking solution that leveraging the influential Big Five personality traits to bridge communication gaps across various personality types. This innovative model integrates with users' emotional states, aided by sentient AI components, enhancing empathy and understanding. By seamlessly adapting its responses to different personality types, the model provides more authentic and meaningful conversations. Furthermore, the integration of sentient AI components amplifies its empathetic capabilities, enabling it to evaluate and respond to subtle emotional states, thereby fostering a deeper understanding of users' mental health, a field that is becoming ever more important.7:205
11
AnishBaghelLiuAn Explainable AI (XAI) Model that Detects Pneumonia From X-Ray ImagesOur project centers on utilizing Explainable AI (XAI) methods on our CNN models to detect pneumonia in X-ray lung images. Leveraging techniques such as LIME, Saliency Maps, Grad-CAM, Occlusion Tests, Deep Taylor Decomposition, and SHAP, we were able to visualize the results of our models. Our XAI-focused approach enhances transparency, enabling insights into the neural network decision process. By combining multiple XAI methods to X-ray images, we successfully distinguished between pneumonia-affected and healthy lung conditions.7:401
12
JuliaBamfordAmerGeopolymer Concrete: Review of Mechanical Properties and Mixture Design in Comparison to Cement Concrete.Concrete, the second most utilized substance globally after water, plays a vital role in infrastructure and daily life. However, its heavy reliance on Portland cement contributes 4-8% of global carbon emissions. This necessitates reducing its environmental impact. Geopolymer concrete, an alternative, utilizes industrial waste and supplementary cementitious materials (SCMs) such as fly ash and glass fibers, presenting potential benefits including a possible 80% reduction in emissions, increased strength, and enhanced durability. Despite these advantages, challenges including reproducibility, the presence of hazardous corrosive activators, and brittle properties persist. This work emphasizes the urgency of addressing carbon emissions by considering geopolymer concrete as a viable partial or full replacement for traditional OPC. While highlighting the benefits, it acknowledges the existing obstacles, such as the limited understanding of optimal mixtures. The research also intends to facilitate comparison across different industrial waste-based geopolymer formulations. Addressing these challenges and engaging with existing research could significantly enhance its contribution to sustainable concrete production.7:006
13
ShivanshBansalMcMahanUsing Quantum Neural Networks (QNNs), Quantum Vision Transformers (QVT), and the Mathematical Morphological Reconstruction Algorithm (MMR) for Brain Tumor DetectionBrain tumors affect millions around the world, so detection is critical to helping doctors determine treatment. Currently, radiologists manually identify tumors through MRI (Magnetic Resonance Imaging) scans; however, this poses several limitations: it creates a heavy reliance on the experience of radiologists, has become increasingly costly and time-consuming, and is not as accessible to areas that lack the necessary resources and doctors. With the advancement of deep learning algorithms, a more accessible and efficient solution is possible. Given the existing research in classical Convolutional Neural Networks (CNNs) for tumor detection, Quantum Convolutional Neural Networks (QCNNs) and Quantum Vision Transformers (QVT) offer a promising approach to the problem. Mathematical Morphological Reconstruction (MMR), another image processing method, provides a relative metric for success in the QCNN, and is another classical alternative to CNNs. This research compares the accuracy and computational speed of the MMR, QCNN, QVT, and CNN algorithms to determine whether introducing a quantum aspect presents any noticeable advantage. To build these models, extensive datasets of MRI brain scans were collected. The MMR algorithm involved applying various techniques such as dilation, erosion, and skull stripping through OpenCV2's morphology functions. The QCNN algorithm utilizes quantum power to encode the data into a parametrized quantum circuit and apply convolutional and pooling layers. In terms of future steps, QVTs will be implemented with QCNNs for higher spatial understanding. So far, our results indicate that the MMR algorithm achieved up to 92% accuracy. These results will be compared with the accuracy of the QCNN, QVT, and CNN algorithms.7:403
14
SaryuBapatlaBrahInvestigating the interaction between clozapine and norclozapine with HLA proteins in the pathogenesis of clozapine-induced agranulocytosisClozapine-induced agranulocytosis (CLIA), is a rare but consequential reaction associated with the clozapine medication for resistant schizophrenia. Various classes of the HLA gene, which helps the immune system detect foreign pathogens, have been found to be associated with an increased risk of CLIA. However, the specific underlying mechanism remains unclear. We postulated that clozapine, or its metabolite norclozapine, may be interacting with HLA proteins to produce the intracellular aberrations that ultimately lead to CLIA. Our research revealed that specific HLA-B protein complexes increase the risk of CLIA by showing a stronger affinity for clozapine and norclozapine, offering new insights into treatment for resistant schizophrenia.7:301
15
ShreyaBaroniaDowningSoft Robotic Gripper for Ocean Trash CleanupWith microplastics and millions of tons of ocean trash entering the ocean each year, marine life has experienced significant repercussions. Plastics and trash that find their way into the oceans kill thousands of animals every year from entanglement and create ocean “dead zones” in which animals cannot survive. Significant progress has to be made to clean the oceans of floating debris to prevent marine life from dying. However, traditional equipment such as nets cannot efficiently collect small pieces of plastic; the use of soft robotics can mitigate this problem. Soft robots can be deployed into the ocean to pick up underwater debris that is out of the range of nets. The proposed mechanisms are a cable-driven gripper modeled on a feather star, a mothership paired with several smaller drones, and a starfish-inspired gripper. The feather star, which will be coated in adhesive, increases the probability of collecting floating trash. The mothership paired with drones is also a viable option due to its ability to cover larger areas through passive motion and pick up a variety of garbage. Each drone will have a unique gripper, making it an efficient solution to the problem at hand. The starfish-inspired gripper is cable-actuated and curves inwards to grip any larger pieces of floating debris, which helps when specific pieces of trash need to be targeted to be picked up.7:405
16
ManasBasavarajuJahanikiaUsing Language Models to Empower Human CreativitySociety often considers AI to be less creative than humans, but our study challenges this belief by exploring the creativity of large language models (LLMs). Focusing on AI’s ability to impersonate individuals, we used established measures such as the Alternate Uses Task (AUT), Remote Associates Task (RAT), and Torrance Tests of Creative Thinking (TTCT) to compile an original creativity assessment that quantifies both convergent and divergent thinking. Our test also includes the NEO Five-Factor Inventory (NEO-FFI), a self-reported personality measure, to generate an extensive profile of the participant’s creativity. To compare the LLM against human participants, we provided the LLM with the participants’ demographic data, including but not limited to race, age, gender, profession, salary, and location; this allowed for more personalized responses from the LLM. Our preliminary data indicates that the LLM demonstrates significant creativity, outscoring humans on all three assessments while offering novel, comprehensive, and relevant responses. Our findings allow for further exploration into the dynamics between humans and AI as a means to bolster human creativity.7:206
17
ShriyaBhamidipatiJahanikiaAI Assist: Exploring the Feasibility of Biofeedback Mechanisms for Managing Symptoms of Attention-Deficit/Hyperactivity DisorderAttention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder characterized by symptoms of inattention, hyperactivity, and impulsivity that can significantly impair academic, social, and occupational functioning. Traditional interventions for ADHD often rely on student-teacher-parent interactions to manage symptoms, but the implementation and effectiveness of these interventions can vary greatly. In recent years, there has been increasing interest in exploring the use biofeedback mechanisms with Natural Language Processing (NLP)-based interventions for managing symptoms of ADHD. This presentation aims to explore the efficacy and feasibility of biofeedback mechanisms for monitoring and managing symptoms of ADHD through the use of wearable sensors, with a focus on the importance of student-teacher-parent interactions in promoting positive outcomes. By examining the current research and identifying potential areas for future investigation, this presentation seeks to provide a comprehensive overview of the potential benefits and limitations of biofeedback mechanisms for managing symptoms of ADHD, and to highlight the importance of collaboration between students, teachers, and parents in supporting individuals with ADHD in order to improve their learning experience.7:107
18
MaitreyiBharathWangEngineering Bispecific Antibodies To Target Acute Myeloid Leukemia CellsAcute myeloid leukemia (AML) is a rare blood and bone marrow cancer that primarily affects adults. Acute myeloid leukemia is one of the most common forms of cancer in adults, with a survival rate that ranges from 10% to 69% depending on age. AML causes excessive production of abnormal white blood cells, red blood cells, and platelets that crowd out healthy cells and increase the risk of infection, anemia, and other blood-related health conditions. In this project, we aim to engineer bispecific antibodies to target antigens presented in acute myeloid leukemia cells. Bispecific antibodies are Y-shaped proteins used to target harmful pathogens in the body. These antibodies have two fragment antibody binding regions (Fabs), allowing them to simultaneously bind to tumor antigen proteins CLL1 or CD123 and to T cells at T cell engager CD3. Through this process, multiple T-cells are recruited toward the tumor and are able to eliminate the AML-affected cells through signaling, proliferation, and redirection of cytotoxic response. There are other therapies as well including biologicals, a class of drug derived from living organisms, which are currently not FDA approved and CAR T-cell therapy which is currently under clinical trials. However, bispecific antibodies are engineered to be able to target specific antigens with improved efficacy and safety. In our project, we produced multiple variations of antibodies through cell transfection of plasmid DNA into HEK293-F mammalian cells (human embryonic kidney cells). In future steps, we plan to purify our antibodies and test their binding affinities to AML antigens.7:503
19
AristaaBhardwajDowning/NjooA Comparative Analysis of Recent Progress in Blind Docking Softwares for Protein-Ligand InteractionsIn this study, we consider the advantage of blind docking as an unbiased way to identify and predict protein structures, binding sites, and their receptors to conduct a comparison of these softwares. With a thorough comparison of various blind-docking software, we aim to identify the most efficient and accurate methodology for blind-docking. We determined the scope of the analysis and comparison as including accuracy, predictive capability, algorithmic methodologies used, and the diversity of the data. Regarding the examination of these factors in an assortment of blind-docking software, each application generates information about the docking of the ligand. Thus, we analyze variables with strong correlation to the protein structures, such as the scoring function with the affinity of the ligand to the binding pocket and the estimated fitness (kcal/mol). The intention of our study is to provide a comprehensive analysis and comparison of blind-docking software for use in research pertaining to drug discovery and protein structure identification.7:505
20
TanushBhardwajDowningDiscovering mini black holes through observations of their stellar companion in binary systemsIn April 2021, Jayasinghe et al. reported the discovery of a mass-gap black hole binary companion to the red giant V723 Mon. Building upon their astronomical models, we developed an adaptable program to uncover potential star and mass-gap black hole binary systems in large-scale datasets. Using data from ESA’s Gaia DR2 and NASA’s exoplanet archive, we employed computational analysis methods to review distortions in the stellar companion’s envelope and shifts in its radial velocity by considering the star’s increasing error in radius measurements and deviations in mean radial velocity, respectively. This resulted in the discovery of 19 potential black hole candidates at a distance of 19.5-1,100 parsecs from Earth. To create a radial velocity time series for each celestial object, we imported datasets from Carmenes, HARPS-N, among other sources. Using R’s libraries (astrochron, ggplot, and dplyr), we were able to graphically represent each star’s radial velocity time series, extracting orbital period. Referencing a mass function from Jayasinghe et al., we then approximated the mass of the potential candidates, providing strong indications that these stars exist in a binary system with a mass-gap black hole of ~3-10☉.7:202
21
AparnaBhaskarCunhaComparative Genomics to Discover Novel Relationships between Sharks and Humans in the Context of Colorectal CancerThis study compares elephant, whale, and white shark genomes with the human genome to identify potential novel tumor-suppressing regions of the shark genome.7:407
22
HiyaBhatiaDeGrendeleEnhancing Computational Fluid Dynamics Through SimulationIn this study, we present Computational Fluid Dynamics (CFD) simulation techniques to advance numerical modeling possibilities. Initial simulations focused on the solution of the advection equation and the Burgers equation, and were expanded upon with 2D models, boundary conditions, and grid-detecting programs. In future work, we will continue to build up our simulation framework to take into account more accurate physics and more complex objects, such as airplane wings. This work, in conjunction with machine learning testing, not only offers faster simulations but also holds the potential to revolutionize various engineering disciplines by offering more efficient and accurate predictive modeling capabilities.7:008
23
NikhilBhowmikAdamsAlgae powered batteryOur group is using Synochocytis algae to power a battery, seeing how long the home grown algae will power the battery compared to past trials7:001
24
ShreyBirmiwalDowningDetermining Trends in Pharmacokinetic Properties in FDA Approved DrugsPharmacokinetics (PK) is the study of a drug and its metabolite kinetics in the body - specifically, it refers to the evaluation of a drug and its metabolites in various areas of the body to determine its efficacy at administering treatment and prevent failures in drug design. At the cornerstone of PK analysis is ADME properties, standing for absorption, distribution, metabolism, and excretion, four core properties which define a drug’s efficacy. Meanwhile, recent advances in deep learning have expanded the role of predictive ADME in medicinal chemistry, giving rise to various PK predictors to analyze the properties of a given drug. This paper aims to compare various web-based PK predictors and use them to determine trends in FDA approved drugs’ active ingredients, as well as use a weighted-score analysis to assign a pharmacokinetic score to each molecule. We aim to provide conclusions on ideal PK properties for drugs for FDA-approval as well as provide a standard for comparison of web-based PK predictors for future work in the field.7:002
25
PragyaaBodapatiJahanikiaCovidVacMap: A Global Network Analysis of COVID-19 Vaccine Distribution to Predict Breakthrough CasesThe ongoing COVID-19 pandemic, also known as the coronavirus pandemic, is a global pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Multiple vaccines have been developed, underwent clinical trials, and are being distributed along with many boosters and doses. While these vaccines are currently being distributed the virus still spreads as we have not reached herd immunity. As the virus spreads new variants enter.7:501
26
SindhujaBokkisamRengananthanComparing the potency of different Polyphenols in the Suppression of Amyloid-Beta Proteins in Alzheimer's disease Using Caenorhabditis Elegans as model organismAlzheimer’s disease (AD), a neurodegenerative disease that leads to cognitive memory decline, is caused by the aggregation of Amyloid-β proteins through the formation of β and γ-secretases. These secretases cut down the Amyloid-β precursor protein responsible for neural growth and repair, and result in abnormal levels of Amyloid-β (Aβ42) proteins. Natural polyphenols have antioxidative and neuroprotective properties that limit the aggregation of Aβ42 and could be used as a treatment for AD. In this study, Caenorhabditis Elegans (C. Elegans) was used as a model organism for AD as it possesses various human homologous genes and protein networks involved in the disease. The in-vivo paralysis assay evaluated the response of the nematodes in an environment containing polyphenols. The lifespan assay, which is currently in progress, will analyze worm survivability in different polyphenolic environments. From these studies, we can evaluate polyphenols with the potential to limit aggregation of Aβ42 proteins.7:402
27
AnavBordiaMcMahanUsing Convolutional Neural Networks to Diagnose Pigmented Skin LesionsSkin cancer poses a significant health challenge, with early diagnosis crucial for effective treatment. Inspection by dermatologists remains the standard for diagnosis, but it can be subjective and prone to error. This project explores the potential of convolutional neural networks (CNNs) as a powerful tool for skin lesion classification, aiming to improve accuracy and accessibility of diagnosis and ultimately contribute to better patient outcomes. Leveraging the data of over 10,000 images of pigmented skin lesions categorized into seven diagnostic classes (including benign and malignant tumors), we will use a deep learning model based on CNNs. This project showcases the power of deep learning and CNNs in revolutionizing skin cancer diagnosis.7:304
28
AmbareeshBudarajuPoudyalExploring the Future of Agriculture: Advances in Plant Tissue Culture for Sustainable GrowthThis study is dedicated to investigating the future of agriculture through the proliferation of selected fruit species, with the goal of optimizing plant regeneration for molecular breeding eventually for sustainable agriculture. The explant of some fruit specimens were initially collected from the field and subsequently proliferated in-vitro by using plant tissue culture technique. Rigorous sterilization protocols and expert culturing practices were employed during the proliferation process to mitigate the risk of microbial contamination in the in-vitro grown plants. Subsequent growth of these plants, following the proliferation, was facilitated by the application of MS Media. This growth medium incorporates a precise blend of phytohormones, organic supplements, and essential components to ensure quality and rapid propagation. These preliminary findings lay the groundwork for future research pathways, particularly in the field of organogenesis and crop regeneration post gene transformation. Ultimately, these findings lead to expanded studies on genetic modified organisms.7:504
29
ShohiniChakrabortyJahanikiaInvestigating the Impact of Meditation on Neural Processes: Insights from EEG and Sleep AnalysisThe popularity in the research of the impact of meditation on brain activity has spiked over the past few years. Even so, analysis on the effects of mediation through electroencephalography (EEG) has been done for decades but its impacts are still uncertain. This is due to how various meditation practices affect brain activity differently, shown evident in our dataset. The dataset included data from four blocks of meditation: two thinking blocks, one breathing block, and one tradition specific meditation block. The first meditation set is a breath count meditation which may contain sleep data. The first thinking task will work as the control data to compare with the breathing task. After sorting through 50 subjects, and running it through our software titled EEGLab, the data had to be pre-processed to make accurate conclusions. With continuing data analysis through EEGLab, and identifying sleep pattern in our data using various filtering and computing methods, we hope to find the direct effects of meditation and how it may benefit cognition, perception, and emotional processing.7:507
30
AdithiChandraCunhaComparative Genomics to Discover Novel Relationships between Sharks and Humans in the Context of Colorectal CancerThis study compares elephant, whale, and white shark genomes with the human genome to identify potential novel tumor-suppressing regions of the shark genome.7:407
31
LindaChangMcMahanGenerating Chemically Stable Molecules via Quantum Computing and Python Molecular Benchmarking ProcessesCurrent drug discovery and development processes can cost upwards to billions of dollars and last five to twelve years to create one FDA-approved drug, so researchers have been implementing computational chemistry methods to current molecular synthesis pathways. However, current computational chemistry methods often have inefficient runtimes, as the chemical space is vast. Aiming for a more efficient runtime and robust analysis of high-dimensional molecular data, our group previously implemented the Hybrid Quantum-Classical Generative Adversarial Network (QGAN) and the most recent iteration of the model is the Hybrid Quantum-Classical Graph Generative Adversarial Network (QNetGAN) to synthesize chemically feasible molecules. The QNetGAN addresses the issues faced by the QGAN including the distance between atoms exceeding bonding length, which causes most of the atoms in the generated molecules to remain unbonded. By generating graphs and utilizing long-short term memory cells, QNetGAN generated 141/300 structurally valid molecules that satisfy Lipinski’s Rule of Five, yielding a 47% success rate — a notable increase from the prior 2.3% — with a minimal training time of 10.164 minutes. However, molecules generated by QNetGAN remain a work in progress, with the majority of molecules failing to satisfy the Octet Rule and having unoptimized bond lengths and bond angles. Our work focuses on implementing chemical post-processing algorithms to increase the chemical structures’ stability and feasibility. Our Python-based Octet Rule algorithm employs a Depth-First Search (DFS) traversal structure to count the number of bonds and open orbitals on each atom. Then, our formal charge calculation algorithm calculates the formal charge of each atom, to find the best overall structure of the molecule with the most stability. Finally, our Hydrogen Addition Algorithm builds upon the Octet Rule algorithm to traverse through the molecule’s adjacency matrix, adding hydrogens to central atoms where necessary to complete the molecule. Although a work in progress, our post-processing algorithms have been able to check for Octet Rule satisfaction, calculate formal charge, and add hydrogen atoms to molecules as necessary with 100% accuracy. In the future, we plan to continue testing our algorithms on larger molecular structures, while implementing methods to handle exceptional cases to the basic rules. Furthermore, as our generated molecules are initially only given 2D coordinates due to software limitations, we are working to recalculate each atom’s coordinates such that the final molecular structure corresponds correctly to their molecular geometries (i.e. tetrahedral, trigonal planar, bent, etc.) and bond angles.7:103
32
DaphneChaoJahanikiaInvestigating the Impact of Meditation on Neural Processes: Insights from EEG and Sleep AnalysisThe popularity in the research of the impact of meditation on brain activity has spiked over the past few years. Even so, analysis on the effects of mediation through electroencephalography (EEG) has been done for decades but its impacts are still uncertain. This is due to how various meditation practices affect brain activity differently, shown evident in our dataset. The dataset included data from four blocks of meditation: two thinking blocks, one breathing block, and one tradition specific meditation block. The first meditation set is a breath count meditation which may contain sleep data. The first thinking task will work as the control data to compare with the breathing task. After sorting through 50 subjects, and running it through our software titled EEGLab, the data had to be pre-processed to make accurate conclusions. With continuing data analysis through EEGLab, and identifying sleep pattern in our data using various filtering and computing methods, we hope to find the direct effects of meditation and how it may benefit cognition, perception, and emotional processing.7:507
33
RishiChapatiAmerGeopolymer Concrete: Review of Mechanical Properties and Mixture Design in Comparison to Cement Concrete.Concrete, the second most utilized substance globally after water, plays a vital role in infrastructure and daily life. However, its heavy reliance on Portland cement contributes 4-8% of global carbon emissions. This necessitates reducing its environmental impact. Geopolymer concrete, an alternative, utilizes industrial waste and supplementary cementitious materials (SCMs) such as fly ash and glass fibers, presenting potential benefits including a possible 80% reduction in emissions, increased strength, and enhanced durability. Despite these advantages, challenges including reproducibility, the presence of hazardous corrosive activators, and brittle properties persist. This work emphasizes the urgency of addressing carbon emissions by considering geopolymer concrete as a viable partial or full replacement for traditional OPC. While highlighting the benefits, it acknowledges the existing obstacles, such as the limited understanding of optimal mixtures. The research also intends to facilitate comparison across different industrial waste-based geopolymer formulations. Addressing these challenges and engaging with existing research could significantly enhance its contribution to sustainable concrete production.7:006
34
AryaChatterjeeJahanikiaDietary patterns during COVID-19: Analyzing changes in consumption habits in adults before, during, and after COVID-19 lockdown"The COVID-19 pandemic and the resulting quarantine have caused drastic changes in an individual’s lifestyle. Dietary lifestyle, relating to an individual’s relationship with both food and exercise, is a major part of this change. We hypothesize that the pandemic caused participants to change their consumption patterns drastically, and that these changes have remained in place even after the lockdown period ended. Our questionnaire, which borrows ideas from four standardized eating assessments (SCOFF, EAT-26, CET, and EDE-Q) will aim to test this in adult participants. After completing data collection, we will conduct data analysis using R and various statistical softwares. The results of this study will provide a better understanding of one of the several changes that the pandemic has brought upon us."7:408
35
NeelChatterjeeJahanikiaDietary patterns during COVID-19: Analyzing changes in consumption habits in adults before, during, and after COVID-19 lockdown"The COVID-19 pandemic and the resulting quarantine have caused drastic changes in an individual’s lifestyle. Dietary lifestyle, relating to an individual’s relationship with both food and exercise, is a major part of this change. We hypothesize that the pandemic caused participants to change their consumption patterns drastically, and that these changes have remained in place even after the lockdown period ended. Our questionnaire, which borrows ideas from four standardized eating assessments (SCOFF, EAT-26, CET, and EDE-Q) will aim to test this in adult participants. After completing data collection, we will conduct data analysis using R and various statistical softwares. The results of this study will provide a better understanding of one of the several changes that the pandemic has brought upon us."7:408
36
AdelinaChauMcMahanGenerating Chemically Stable Molecules via Quantum Computing and Python Molecular Benchmarking ProcessesCurrent drug discovery and development processes can cost upwards to billions of dollars and last five to twelve years to create one FDA-approved drug, so researchers have been implementing computational chemistry methods to current molecular synthesis pathways. However, current computational chemistry methods often have inefficient runtimes, as the chemical space is vast. Aiming for a more efficient runtime and robust analysis of high-dimensional molecular data, our group previously implemented the Hybrid Quantum-Classical Generative Adversarial Network (QGAN) and the most recent iteration of the model is the Hybrid Quantum-Classical Graph Generative Adversarial Network (QNetGAN) to synthesize chemically feasible molecules. The QNetGAN addresses the issues faced by the QGAN including the distance between atoms exceeding bonding length, which causes most of the atoms in the generated molecules to remain unbonded. By generating graphs and utilizing long-short term memory cells, QNetGAN generated 141/300 structurally valid molecules that satisfy Lipinski’s Rule of Five, yielding a 47% success rate — a notable increase from the prior 2.3% — with a minimal training time of 10.164 minutes. However, molecules generated by QNetGAN remain a work in progress, with the majority of molecules failing to satisfy the Octet Rule and having unoptimized bond lengths and bond angles. Our work focuses on implementing chemical post-processing algorithms to increase the chemical structures’ stability and feasibility. Our Python-based Octet Rule algorithm employs a Depth-First Search (DFS) traversal structure to count the number of bonds and open orbitals on each atom. Then, our formal charge calculation algorithm calculates the formal charge of each atom, to find the best overall structure of the molecule with the most stability. Finally, our Hydrogen Addition Algorithm builds upon the Octet Rule algorithm to traverse through the molecule’s adjacency matrix, adding hydrogens to central atoms where necessary to complete the molecule. Although a work in progress, our post-processing algorithms have been able to check for Octet Rule satisfaction, calculate formal charge, and add hydrogen atoms to molecules as necessary with 100% accuracy. In the future, we plan to continue testing our algorithms on larger molecular structures, while implementing methods to handle exceptional cases to the basic rules. Furthermore, as our generated molecules are initially only given 2D coordinates due to software limitations, we are working to recalculate each atom’s coordinates such that the final molecular structure corresponds correctly to their molecular geometries (i.e. tetrahedral, trigonal planar, bent, etc.) and bond angles.7:103
37
MichaelCheklerMcMahanExploring Potential Solutions To The Small-Scale Problems of the Lambda Cold Dark Matter ModelThe Lambda Cold Dark Matter (λCDM), which depicts dark matter particles as cold and collisionless, is successful for large scale structures but has faced challenges on the small scale due to contradictions with astronomical observations. The challenges faced are called the small-scale problems, which are: the core-cusp problem, the missing satellites problem, and the too-big-to-fail problem. Alternate models, such as the Self Interacting Dark Matter (SIDM) model or the implementation of baryonic feedback effects, have been proposed to combat these small scale problems. Unlike λCDM particles that primarily interact through gravity, SIDM particles do interact with each other through means of self-scattering processes. In this paper, we utilize recent simulations to explore the possibility of the SIDM model with baryonic feedback combatting the small-scale problems while accurately representing the universe on the large scale to ultimately replace the λCDM model.7:204
38
EmilyChenCunhaComparative Genomics to Discover Novel Relationships between Sharks and Humans in the Context of Colorectal CancerThis study compares elephant, whale, and white shark genomes with the human genome to identify potential novel tumor-suppressing regions of the shark genome.7:407
39
HarrietChenNjooC-4 analogs of podophyllotoxin as tubulin inhibitors: Synthesis, biological evaluation, and structure-activity relationshipThe diversity of lignin small molecules derived from podophyllotoxin, a non-covalent tubulin inhibitor isolated from the Podophyllum family, has led to the clinical development of FDA-approved anticancer agents etoposide and teniposide. While these two compounds share the same tetracyclic core as podophyllotoxin, two subtle structural changes—4’ methylation on the aromatic ring and stereospecific glycosylation at the C-4 hydroxyl—result in an alternate biological mechanism. Given the immense pharmacological importance of these two features, we synthesized and evaluated a systematic library of diversified esters to establish a structure-activity relationship regarding modification at C-4 on the properties of podophyllotoxin. We determined the biological activity of these esters through cell viability assays, computer docking models, tubulin polymerization assays, and cell cycle analysis. Altogether, we demonstrate that increasing steric hindrance at C-4 leads to a loss in potency against human cancer cells but has a significantly lesser impact on cell-free tubulin inhibition. This suggests that the biological activity of our compounds may be attributed to cytosolic and membrane distribution rather than binding interactions within the colchicine site.7:208
40
SeanChenDowningApplying Supervised and Unsupervised Machine Learning Techniques to Characterize Habitability in ExoplanetsFor decades, researchers have searched the night sky for habitable exoplanets. The main criterion for an exoplanet’s habitability involves its ability to host liquid water, which depends on a multitude of characteristics unique to each exoplanet and its host star, all of which must work together to create an environment capable of sustaining life. We first deduplicated the open-source NASA Exoplanet Archive dataset for efficient calculations, creating one cumulative entry for every exoplanet. Then, we calculated each host star’s bolometric luminosity and determined whether each exoplanet resided within the bounds of their host stars’ circumstellar habitable zone (CHZ). This involved attributes such as pl_orbsmax, pl_orbeccen, and st_lum. Another method in the pursuit of a supervised learning approach involved the examination of high-value candidates [e.g.: K2 18 b & K2 3 d] from prior research in regards to additional explanatory variables that might refine our algorithm's ruleset. Finally, an unsupervised model was also used to find similarities between planets and to cluster them by attribute. The data were first de-duplicated such that every measurement of a planet was averaged, after which the remaining data was run through the algorithm and represented in a 3D map. For confirmation, we compared the results of these three methods. The CHZ algorithm, in particular, corroborated two candidates with the NASA Planetary Habitability Laboratory (PHL): Kepler-62 f and Kepler-1229 b, which exhibited significant values for major explanatory variables.7:102
41
CherylCheungCunhaComparative Genomic Analysis of Colorectal Cancer Microbiome Bacteria to Discover Novel RelationshipsColorectal cancer (CRC) is uncontrolled tumor growth that starts in the rectum or colon (Park E. et al., 2022). Many factors affect the development of cancer, including daily habits, environments, and genetics. Our research focuses on analyzing the differences in pathways/enzymes between cancerous and non-cancerous associated bacteria in the gut microbiome outlined by a recent cancer microbiome review (Park E. et al., 2022). By utilizing the Bacterial and Viral Bioinformatics Resource Center (BV-BRC), we compiled our bacteria’s genetic information into genome groups and used the comparative systems service to identify target pathways and construct phylogenetic trees. After focusing on genomes, we delved deeper into the enzymes. The programming language R was used to narrow down four specific enzymes from the set of genomes: two from the pathways only in non-cancerous bacteria and two in cancerous-associated bacteria. A Multiple Sequence Alignment (MSA) run at the genome level identified the range of lowest entropy among the genes in the four enzymes - one of which had the lowest range of 30-40. We are using NCBI Blast and other bioinformatics methods to characterize/validate the four enzymes in our respective target bacteria. Our end goal is to target/screen the unique pathways and enzymes (like the enzyme with EC number 5.4.3.2) of the cancer-associated bacteria and non-cancerous associated bacteria to decrease the metastasis of CRC tumors (Park E. et al., 2022). These genes, that help create the enzymes, can be manipulated in the wet lab as shown by the cited paper.7:307
42
FalakChhatreMcMahanGenerating Chemically Stable Molecules via Quantum Computing and Python Molecular Benchmarking ProcessesCurrent drug discovery and development processes can cost upwards to billions of dollars and last five to twelve years to create one FDA-approved drug, so researchers have been implementing computational chemistry methods to current molecular synthesis pathways. However, current computational chemistry methods often have inefficient runtimes, as the chemical space is vast. Aiming for a more efficient runtime and robust analysis of high-dimensional molecular data, our group previously implemented the Hybrid Quantum-Classical Generative Adversarial Network (QGAN) and the most recent iteration of the model is the Hybrid Quantum-Classical Graph Generative Adversarial Network (QNetGAN) to synthesize chemically feasible molecules. The QNetGAN addresses the issues faced by the QGAN including the distance between atoms exceeding bonding length, which causes most of the atoms in the generated molecules to remain unbonded. By generating graphs and utilizing long-short term memory cells, QNetGAN generated 141/300 structurally valid molecules that satisfy Lipinski’s Rule of Five, yielding a 47% success rate — a notable increase from the prior 2.3% — with a minimal training time of 10.164 minutes. However, molecules generated by QNetGAN remain a work in progress, with the majority of molecules failing to satisfy the Octet Rule and having unoptimized bond lengths and bond angles. Our work focuses on implementing chemical post-processing algorithms to increase the chemical structures’ stability and feasibility. Our Python-based Octet Rule algorithm employs a Depth-First Search (DFS) traversal structure to count the number of bonds and open orbitals on each atom. Then, our formal charge calculation algorithm calculates the formal charge of each atom, to find the best overall structure of the molecule with the most stability. Finally, our Hydrogen Addition Algorithm builds upon the Octet Rule algorithm to traverse through the molecule’s adjacency matrix, adding hydrogens to central atoms where necessary to complete the molecule. Although a work in progress, our post-processing algorithms have been able to check for Octet Rule satisfaction, calculate formal charge, and add hydrogen atoms to molecules as necessary with 100% accuracy. In the future, we plan to continue testing our algorithms on larger molecular structures, while implementing methods to handle exceptional cases to the basic rules. Furthermore, as our generated molecules are initially only given 2D coordinates due to software limitations, we are working to recalculate each atom’s coordinates such that the final molecular structure corresponds correctly to their molecular geometries (i.e. tetrahedral, trigonal planar, bent, etc.) and bond angles.7:103
43
ArfathChowdhuryJahanikiaInvestigating the Impact of Meditation on Neural Processes: Insights from EEG and Sleep AnalysisThe popularity in the research of the impact of meditation on brain activity has spiked over the past few years. Even so, analysis on the effects of mediation through electroencephalography (EEG) has been done for decades but its impacts are still uncertain. This is due to how various meditation practices affect brain activity differently, shown evident in our dataset. The dataset included data from four blocks of meditation: two thinking blocks, one breathing block, and one tradition specific meditation block. The first meditation set is a breath count meditation which may contain sleep data. The first thinking task will work as the control data to compare with the breathing task. After sorting through 50 subjects, and running it through our software titled EEGLab, the data had to be pre-processed to make accurate conclusions. With continuing data analysis through EEGLab, and identifying sleep pattern in our data using various filtering and computing methods, we hope to find the direct effects of meditation and how it may benefit cognition, perception, and emotional processing.7:507
44
AndrewChyuNjooSynthesis of Dexamethasone and Related Fluorinated Corticosteroid ProdrugsSince the initial success of fluorinated corticosteroids in the 1950s, several, including dexamethasone, triamcinolone acetonide, and betamethasone, have been investigated for their potent anti-inflammatory activities and improved pharmacokinetic profiles. Among these, dexamethasone received FDA approval in 1958 and has been prescribed to millions for the treatment of inflammation, arthritis, asthma, and multiple sclerosis. Prednisolone, a similar corticosteroid, is delivered as its oxidized prodrug, prednisone, to improve its metabolic properties with a longer half-life. Inspired by this, we attempted a similar synthetic strategy on dexamethasone. We hypothesized the oxidation of its secondary alcohol at C-11 into a ketone might be an effective prodrugging strategy, as 11β-Hydroxysteroid dehydrogenase will dehydrogenate the ketone back into an alcohol. In our work, we investigate various analogs comparing dexamethasone to other fluorinated corticosteroids to evaluate the quality of prodrugging work.7:108
45
AthenaContosPoudyalExploring the Future of Agriculture: Advances in Plant Tissue Culture for Sustainable GrowthThis study is dedicated to investigating the future of agriculture through the proliferation of selected fruit species, with the goal of optimizing plant regeneration for molecular breeding eventually for sustainable agriculture. The explant of some fruit specimens were initially collected from the field and subsequently proliferated in-vitro by using plant tissue culture technique. Rigorous sterilization protocols and expert culturing practices were employed during the proliferation process to mitigate the risk of microbial contamination in the in-vitro grown plants. Subsequent growth of these plants, following the proliferation, was facilitated by the application of MS Media. This growth medium incorporates a precise blend of phytohormones, organic supplements, and essential components to ensure quality and rapid propagation. These preliminary findings lay the groundwork for future research pathways, particularly in the field of organogenesis and crop regeneration post gene transformation. Ultimately, these findings lead to expanded studies on genetic modified organisms.7:504
46
MaxCuiMcMahanGenerating Chemically Stable Molecules via Quantum Computing and Python Molecular Benchmarking ProcessesCurrent drug discovery and development processes can cost upwards to billions of dollars and last five to twelve years to create one FDA-approved drug, so researchers have been implementing computational chemistry methods to current molecular synthesis pathways. However, current computational chemistry methods often have inefficient runtimes, as the chemical space is vast. Aiming for a more efficient runtime and robust analysis of high-dimensional molecular data, our group previously implemented the Hybrid Quantum-Classical Generative Adversarial Network (QGAN) and the most recent iteration of the model is the Hybrid Quantum-Classical Graph Generative Adversarial Network (QNetGAN) to synthesize chemically feasible molecules. The QNetGAN addresses the issues faced by the QGAN including the distance between atoms exceeding bonding length, which causes most of the atoms in the generated molecules to remain unbonded. By generating graphs and utilizing long-short term memory cells, QNetGAN generated 141/300 structurally valid molecules that satisfy Lipinski’s Rule of Five, yielding a 47% success rate — a notable increase from the prior 2.3% — with a minimal training time of 10.164 minutes. However, molecules generated by QNetGAN remain a work in progress, with the majority of molecules failing to satisfy the Octet Rule and having unoptimized bond lengths and bond angles. Our work focuses on implementing chemical post-processing algorithms to increase the chemical structures’ stability and feasibility. Our Python-based Octet Rule algorithm employs a Depth-First Search (DFS) traversal structure to count the number of bonds and open orbitals on each atom. Then, our formal charge calculation algorithm calculates the formal charge of each atom, to find the best overall structure of the molecule with the most stability. Finally, our Hydrogen Addition Algorithm builds upon the Octet Rule algorithm to traverse through the molecule’s adjacency matrix, adding hydrogens to central atoms where necessary to complete the molecule. Although a work in progress, our post-processing algorithms have been able to check for Octet Rule satisfaction, calculate formal charge, and add hydrogen atoms to molecules as necessary with 100% accuracy. In the future, we plan to continue testing our algorithms on larger molecular structures, while implementing methods to handle exceptional cases to the basic rules. Furthermore, as our generated molecules are initially only given 2D coordinates due to software limitations, we are working to recalculate each atom’s coordinates such that the final molecular structure corresponds correctly to their molecular geometries (i.e. tetrahedral, trigonal planar, bent, etc.) and bond angles.7:103
47
ChristinaDangAmerGeopolymer Concrete: Review of Mechanical Properties and Mixture Design in Comparison to Cement Concrete.Concrete, the second most utilized substance globally after water, plays a vital role in infrastructure and daily life. However, its heavy reliance on Portland cement contributes 4-8% of global carbon emissions. This necessitates reducing its environmental impact. Geopolymer concrete, an alternative, utilizes industrial waste and supplementary cementitious materials (SCMs) such as fly ash and glass fibers, presenting potential benefits including a possible 80% reduction in emissions, increased strength, and enhanced durability. Despite these advantages, challenges including reproducibility, the presence of hazardous corrosive activators, and brittle properties persist. This work emphasizes the urgency of addressing carbon emissions by considering geopolymer concrete as a viable partial or full replacement for traditional OPC. While highlighting the benefits, it acknowledges the existing obstacles, such as the limited understanding of optimal mixtures. The research also intends to facilitate comparison across different industrial waste-based geopolymer formulations. Addressing these challenges and engaging with existing research could significantly enhance its contribution to sustainable concrete production.7:006
48
KavyaDattAmadiDevelopment of synthetic aptamers for use as low-cost PLD1 inhibitorsPLD1 is a gene in the genome of the cell that codes for an enzyme that breaks down phosphatidylcholine into phosphatidic acid and choline. The phosphatidic acid then leaves the cell and attaches to an mTOR receptor in another cell and causes a signal to be sent into the cell. As a result of the signal transduction and a phosphorylation cascade, the cell starts performing mitosis. Due to excessive amounts of phosphatidic acid the mTOR receptor constantly signals to the cell to start mitosis through DAG kinases. PLD1 amplifies anti-apoptotic functions that cause inflammation and create a roadblock for chemotherapy. The focus of our project is to create synthetic aptamers using specialized DNA or RNA bases to act as competitive inhibitors for PLD1 allowing patients to continue chemotherapy.7:101
49
AdityaDawarJahanikiaPsyGPT.ai: Developing an accessible, online source of comprehensive mental health services.Currently, receiving a diagnosis and/or getting therapy for a mental health disorder is not extremely accessible and generally requires multiple doctor visits. We hope to make getting fast and effective therapy easier and more accessible for those who are unable to or don’t feel comfortable visiting a therapist/doctor. Notably, veterans are at a significantly high risk for certain disorders, but mental health resources for this population are still lacking and relatively slow despite the vital importance of swift treatment. Our product aims to assist with the effects of and eventual diagnosis of depression and bipolar disorder. Though the two seem similar, their medical treatments and prognoses are extremely different. We hope to provide a user-friendly, interactive location for anyone to easily access available online mental health resources, such as the suicide and crisis hotline and various online tests.To achieve our goal of creating a successful AI powered-website, we need to create a proper design and develop the API, or, Application Programming Interface. We utilized Figma, a cloud-based design and prototyping tool to produce a user-friendly, approachable design for our chatbot. It is crucial to also design the main frame using HTML and CSS. With the help of Bootstrap, we hope to understand and construct a user-friendly website that can accommodate ChatGPT’s API keys. Once we create our website, we can import ChatGPT and train/test our data, developing a finished product.7:306
50
AdityaDawarJahanikiaUsing Topological Data Analysis (TDA) and Mapper to Study the Homology of Neurodegenerative DiseasesNeurodegenerative diseases disrupt neuronal function, drastically impairing regular bodily activity and threatening the lives of patients. Along with environmental links, the presence of these diseases is often correlated with the expression of certain genes. For example, there is an association between the presence of the e4 form of the APOE gene and the risk of Alzheimer’s disease. While correlations for individual diseases have been pinpointed, further research is needed to identify shared mechanisms between the gene expressions across different diseases. Topological Data Analysis (TDA)—a technique to understand the underlying structure of a dataset—can be used to further analyze these similarities. TDA’s ability to cluster high dimensional data by connecting related data points and revealing the data’s global structure makes it a robust tool to uncover the aforementioned relationships across multiple neurodegenerative diseases. By identifying these clustered subgroups in gene expression using dimensionality reduction algorithms and Mapper, pharmaceutical drugs can be specialized and targeted for patients with neurodegenerative diseases.7:005
51
VictoriaDengJahanikiaPsyGPT.ai: Developing an accessible, online source of comprehensive mental health services.Currently, receiving a diagnosis and/or getting therapy for a mental health disorder is not extremely accessible and generally requires multiple doctor visits. We hope to make getting fast and effective therapy easier and more accessible for those who are unable to or don’t feel comfortable visiting a therapist/doctor. Notably, veterans are at a significantly high risk for certain disorders, but mental health resources for this population are still lacking and relatively slow despite the vital importance of swift treatment. Our product aims to assist with the effects of and eventual diagnosis of depression and bipolar disorder. Though the two seem similar, their medical treatments and prognoses are extremely different. We hope to provide a user-friendly, interactive location for anyone to easily access available online mental health resources, such as the suicide and crisis hotline and various online tests.To achieve our goal of creating a successful AI powered-website, we need to create a proper design and develop the API, or, Application Programming Interface. We utilized Figma, a cloud-based design and prototyping tool to produce a user-friendly, approachable design for our chatbot. It is crucial to also design the main frame using HTML and CSS. With the help of Bootstrap, we hope to understand and construct a user-friendly website that can accommodate ChatGPT’s API keys. Once we create our website, we can import ChatGPT and train/test our data, developing a finished product.7:306
52
ArshiaDesarkarNjooSynthesis of Dexamethasone and Related Fluorinated Corticosteroid ProdrugsSince the initial success of fluorinated corticosteroids in the 1950s, several, including dexamethasone, triamcinolone acetonide, and betamethasone, have been investigated for their potent anti-inflammatory activities and improved pharmacokinetic profiles. Among these, dexamethasone received FDA approval in 1958 and has been prescribed to millions for the treatment of inflammation, arthritis, asthma, and multiple sclerosis. Prednisolone, a similar corticosteroid, is delivered as its oxidized prodrug, prednisone, to improve its metabolic properties with a longer half-life. Inspired by this, we attempted a similar synthetic strategy on dexamethasone. We hypothesized the oxidation of its secondary alcohol at C-11 into a ketone might be an effective prodrugging strategy, as 11β-Hydroxysteroid dehydrogenase will dehydrogenate the ketone back into an alcohol. In our work, we investigate various analogs comparing dexamethasone to other fluorinated corticosteroids to evaluate the quality of prodrugging work.7:108
53
ManushriDhanakotiJahanikiaDecoding Inner Speech with Brain-Computer Interfaces: A Study in EEG Analysis and Machine LearningBrain-Computer Interfaces (BCIs) identify brain signals and convert them into instructions executed by external devices. BCIs have the potential to significantly enhance the quality of life for individuals afflicted with neuromuscular disorders by restoring lost functionality. These disorders often hinder communication abilities, necessitating BCIs capable of deciphering internal speech. Electroencephalography (EEG) is a commonly used noninvasive neuroimaging technique that measures the brain's electrophysiological responses resulting from synchronized neurons. Recent advancements in machine learning have facilitated the detection of brain patterns in EEG data, leading to more promising and dependable BCIs. In this project, we employ a dataset comprising EEG data containing inner speech commands from 10 participants. Through data analysis and the application of machine learning, our objective is to create a model that can accurately interpret inner speech.7:106
54
VishruthDineshJahanikiaUsing Topological Data Analysis (TDA) and Mapper to Study the Homology of Neurodegenerative DiseasesNeurodegenerative diseases disrupt neuronal function, drastically impairing regular bodily activity and threatening the lives of patients. Along with environmental links, the presence of these diseases is often correlated with the expression of certain genes. For example, there is an association between the presence of the e4 form of the APOE gene and the risk of Alzheimer’s disease. While correlations for individual diseases have been pinpointed, further research is needed to identify shared mechanisms between the gene expressions across different diseases. Topological Data Analysis (TDA)—a technique to understand the underlying structure of a dataset—can be used to further analyze these similarities. TDA’s ability to cluster high dimensional data by connecting related data points and revealing the data’s global structure makes it a robust tool to uncover the aforementioned relationships across multiple neurodegenerative diseases. By identifying these clustered subgroups in gene expression using dimensionality reduction algorithms and Mapper, pharmaceutical drugs can be specialized and targeted for patients with neurodegenerative diseases.7:005
55
PhoenixDoDeGrendeleEnhancing Computational Fluid Dynamics Through SimulationIn this study, we present Computational Fluid Dynamics (CFD) simulation techniques to advance numerical modeling possibilities. Initial simulations focused on the solution of the advection equation and the Burgers equation, and were expanded upon with 2D models, boundary conditions, and grid-detecting programs. In future work, we will continue to build up our simulation framework to take into account more accurate physics and more complex objects, such as airplane wings. This work, in conjunction with machine learning testing, not only offers faster simulations but also holds the potential to revolutionize various engineering disciplines by offering more efficient and accurate predictive modeling capabilities.8:008
56
AidenDuggalJahanikiaInvestigating the link between the personality traits of creativity, functional fixedness, and the Big C/Small C continuum through the development and testing of novel creativity tasksFunctional fixedness is a cognitive bias that limits one's ability to think of novel and creative uses for an object, and it is known to limit one's divergent thinking. This study models the relationship between functional fixedness, creativity, and personality types using novel task-based assessments inspired by Duncker's candle problem, the Alternate Uses Test (AUT), the Remote Associates Test (RAT), the ICAA creative achievement assessment, and the BFI-10 personality assessment. As the current study is online via Zoom, an additional measure of rumination was incorporated to examine its potential impact on creativity. This project is currently in the process of analyzing the results and formulating a conclusion.7:508
57
AndrewDuvalMcMahanAdvancing Environmental Mapping and Forest Health Assessments: Integrating Machine Learning Algorithms in Autonomous DronesWith the increasing concern for environmental conservation, there is a growing need for efficient methods of environmental mapping and forest health assessments. However, traditional methods employed by the U.S Forest Health Monitoring have faced controversy due to limited spatial resolution and integration of modern technologies. This research paper explores the application of machine learning algorithms in autonomous drones to conduct forest health assessments. Autonomous drones have the ability to collect timely, up-to-date data, which offers enhanced accuracy. This study focuses on training Deep Learning (DL) models to classify different environmental features based on aerial imagery captured by drones. To achieve accurate and efficient data collection, we will utilize Red-Green-Blue imaging and Convolutional Neural Networks (CNN) with the appropriate evaluation metrics, such as the Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI), and foliage color, to create tree classes and identify forest health indicators. By integrating machine learning algorithms into forest health assessment, this study provides a more efficient, accurate, and up-to-date approach to monitor and evaluate the well-being of forests—supporting ongoing efforts towards environmental management and conservation.7:003
58
BhavyaDwivediSubramanianUsing AI to Predict Stock TrendsWe are creating an AI model that can predict stock trends and fluctuations. We want to start by creating a model and feeding it data on Netflix's stock, and then give it data from other stocks so that eventually it can predict the future stock market. We can accomplish this by using softwares such as ARIMA and Prophet, and testing using RMSE.7:303
59
DivyaEashwerAdamsAlgae powered batteryOur group is using Synochocytis algae to power a battery, seeing how long the home grown algae will power the battery compared to past trials7:001
60
AksithiEswaranCunhaA programming package that uses current computational algorithms to predict S/MARS motifsUsing R, machine learning, and heuristics to develop a highly accurate prediction algorithm to identify S/MARS motifs. S/MARS contribute to chromatin organization and influence various metastatic processes; therefore, they are involved in colorectal cancer.7:207
61
AksithiEswaranCunhaComparative Genomics to Discover Novel Relationships between Sharks and Humans in the Context of Colorectal CancerThis study compares elephant, whale, and white shark genomes with the human genome to identify potential novel tumor-suppressing regions of the shark genome.7:407
62
VincentFuLiuAn Explainable AI (XAI) Model that Detects Pneumonia From X-Ray ImagesOur project centers on utilizing Explainable AI (XAI) methods on our CNN models to detect pneumonia in X-ray lung images. Leveraging techniques such as LIME, Saliency Maps, Grad-CAM, Occlusion Tests, Deep Taylor Decomposition, and SHAP, we were able to visualize the results of our models. Our XAI-focused approach enhances transparency, enabling insights into the neural network decision process. By combining multiple XAI methods to X-ray images, we successfully distinguished between pneumonia-affected and healthy lung conditions.7:401
63
HarshitaGabriCunhaA programming package that uses current computational algorithms to predict S/MARS motifsUsing R, machine learning, and heuristics to develop a highly accurate prediction algorithm to identify S/MARS motifs. S/MARS contribute to chromatin organization and influence various metastatic processes; therefore, they are involved in colorectal cancer.7:207
64
EeshaGadekarlaMcMahanUsing Quantum Neural Networks (QNNs), Quantum Vision Transformers (QVT), and the Mathematical Morphological Reconstruction Algorithm (MMR) for Brain Tumor DetectionBrain tumors affect millions around the world, so detection is critical to helping doctors determine treatment. Currently, radiologists manually identify tumors through MRI (Magnetic Resonance Imaging) scans; however, this poses several limitations: it creates a heavy reliance on the experience of radiologists, has become increasingly costly and time-consuming, and is not as accessible to areas that lack the necessary resources and doctors. With the advancement of deep learning algorithms, a more accessible and efficient solution is possible. Given the existing research in classical Convolutional Neural Networks (CNNs) for tumor detection, Quantum Convolutional Neural Networks (QCNNs) and Quantum Vision Transformers (QVT) offer a promising approach to the problem. Mathematical Morphological Reconstruction (MMR), another image processing method, provides a relative metric for success in the QCNN, and is another classical alternative to CNNs. This research compares the accuracy and computational speed of the MMR, QCNN, QVT, and CNN algorithms to determine whether introducing a quantum aspect presents any noticeable advantage. To build these models, extensive datasets of MRI brain scans were collected. The MMR algorithm involved applying various techniques such as dilation, erosion, and skull stripping through OpenCV2's morphology functions. The QCNN algorithm utilizes quantum power to encode the data into a parametrized quantum circuit and apply convolutional and pooling layers. In terms of future steps, QVTs will be implemented with QCNNs for higher spatial understanding. So far, our results indicate that the MMR algorithm achieved up to 92% accuracy. These results will be compared with the accuracy of the QCNN, QVT, and CNN algorithms.7:403
65
SnehaGadekarlaJahanikiaDietary patterns during COVID-19: Analyzing changes in consumption habits in adults before, during, and after COVID-19 lockdown"The COVID-19 pandemic and the resulting quarantine have caused drastic changes in an individual’s lifestyle. Dietary lifestyle, relating to an individual’s relationship with both food and exercise, is a major part of this change. We hypothesize that the pandemic caused participants to change their consumption patterns drastically, and that these changes have remained in place even after the lockdown period ended. Our questionnaire, which borrows ideas from four standardized eating assessments (SCOFF, EAT-26, CET, and EDE-Q) will aim to test this in adult participants. After completing data collection, we will conduct data analysis using R and various statistical softwares. The results of this study will provide a better understanding of one of the several changes that the pandemic has brought upon us."7:408
66
JaanviGanapathyBrahDocking Ligands using SwissDock and Autodock to learn about the most mutable ligands in PI3KThe purpose of this experiment is to find out which ligands are the most easily mutable which we can tell by checking the ligand binding affinities using Swissdock and Autodock. Of which Maestro was used as a modelling software to analyze physical characteristics to look at for future ligands to test.8:001
67
KrishGaneshJahanikiaDecoding Inner Speech with Brain-Computer Interfaces: A Study in EEG Analysis and Machine LearningBrain-Computer Interfaces (BCIs) identify brain signals and convert them into instructions executed by external devices. BCIs have the potential to significantly enhance the quality of life for individuals afflicted with neuromuscular disorders by restoring lost functionality. These disorders often hinder communication abilities, necessitating BCIs capable of deciphering internal speech. Electroencephalography (EEG) is a commonly used noninvasive neuroimaging technique that measures the brain's electrophysiological responses resulting from synchronized neurons. Recent advancements in machine learning have facilitated the detection of brain patterns in EEG data, leading to more promising and dependable BCIs. In this project, we employ a dataset comprising EEG data containing inner speech commands from 10 participants. Through data analysis and the application of machine learning, our objective is to create a model that can accurately interpret inner speech.7:106
68
KarenGaoMcMahanUsing Convolutional Neural Networks to Diagnose Pigmented Skin LesionsSkin cancer poses a significant health challenge, with early diagnosis crucial for effective treatment. Inspection by dermatologists remains the standard for diagnosis, but it can be subjective and prone to error. This project explores the potential of convolutional neural networks (CNNs) as a powerful tool for skin lesion classification, aiming to improve accuracy and accessibility of diagnosis and ultimately contribute to better patient outcomes. Leveraging the data of over 10,000 images of pigmented skin lesions categorized into seven diagnostic classes (including benign and malignant tumors), we will use a deep learning model based on CNNs. This project showcases the power of deep learning and CNNs in revolutionizing skin cancer diagnosis.7:304
69
DivyaGillAmadiDevelopment of synthetic aptamers for use as low-cost PLD1 inhibitorsPLD1 is a gene in the genome of the cell that codes for an enzyme that breaks down phosphatidylcholine into phosphatidic acid and choline. The phosphatidic acid then leaves the cell and attaches to an mTOR receptor in another cell and causes a signal to be sent into the cell. As a result of the signal transduction and a phosphorylation cascade, the cell starts performing mitosis. Due to excessive amounts of phosphatidic acid the mTOR receptor constantly signals to the cell to start mitosis through DAG kinases. PLD1 amplifies anti-apoptotic functions that cause inflammation and create a roadblock for chemotherapy. The focus of our project is to create synthetic aptamers using specialized DNA or RNA bases to act as competitive inhibitors for PLD1 allowing patients to continue chemotherapy.7:101
70
AbiirGodboleDowningApplying Supervised and Unsupervised Machine Learning Techniques to Characterize Habitability in ExoplanetsFor decades, researchers have searched the night sky for habitable exoplanets. The main criterion for an exoplanet’s habitability involves its ability to host liquid water, which depends on a multitude of characteristics unique to each exoplanet and its host star, all of which must work together to create an environment capable of sustaining life. We first deduplicated the open-source NASA Exoplanet Archive dataset for efficient calculations, creating one cumulative entry for every exoplanet. Then, we calculated each host star’s bolometric luminosity and determined whether each exoplanet resided within the bounds of their host stars’ circumstellar habitable zone (CHZ). This involved attributes such as pl_orbsmax, pl_orbeccen, and st_lum. Another method in the pursuit of a supervised learning approach involved the examination of high-value candidates [e.g.: K2 18 b & K2 3 d] from prior research in regards to additional explanatory variables that might refine our algorithm's ruleset. Finally, an unsupervised model was also used to find similarities between planets and to cluster them by attribute. The data were first de-duplicated such that every measurement of a planet was averaged, after which the remaining data was run through the algorithm and represented in a 3D map. For confirmation, we compared the results of these three methods. The CHZ algorithm, in particular, corroborated two candidates with the NASA Planetary Habitability Laboratory (PHL): Kepler-62 f and Kepler-1229 b, which exhibited significant values for major explanatory variables.7:102
71
IshaanGoelMcMahanA Comparison of Supervised Learning and Deep Reinforcement Learning for Autonomous DrivingAutonomous vehicles are a concept and field of vast interest, such as in the field of robotics engineering and in industries such as automobiles. MLAV1's objective is to identify the best machine learning method (supervised learning v.s. deep reinforcement learning) to train autonomous vehicles and thus achieve greater car safety. To compare the machine learning models’ performance, we built a miniature car (using raspberry pi and various sensors) modeled after an autonomous vehicle as well as a maze and obstacles. The machine learning method that leads to the car completing the maze the fastest, will be identified as the more efficient and accurate.7:104
72
MeherGopalaniAdamsAlgae powered batteryOur group is using Synochocytis algae to power a battery, seeing how long the home grown algae will power the battery compared to past trials7:001
73
AidanGorJahanikiaEnhancing Cognition and Working Memory through a Multidimensional Dopaminergic N-Back Memory GameN-back tasks are a form of cognitive training that requires patients to recall information from a certain stimulus previously shown to them. Cognitive research patients often complain that these tasks are boring and mundane when involved in a study. In our study, we aim to use a Dual N-Back Working Memory (DNB-WM) task to enhance working memory. Through fMRI scanning, it has been proven that N-back tasks improve working memory, an executive function of the brain associated with the prefrontal cortex, frontoparietal network, and salience network. Moreover, it is hypothesized that if dopaminergic pathways, such as the Mesolimbic, Mesocortical, Nigrostriatal, and Tuberoinfundibular, are targeted, dopamine production will increase engagement and productivity during cognitive tests, therefore further increasing working memory.
Our study is designed to involve 32 participants; 16 for an experimental group and 16 for a control group. Patients will be given the list sorting test from the NIH toolbox cognition battery, which will act as a baseline for their working memory. Then, the experimental group will be given a multidimensional N-back game that reflects a ‘gamified’ task, while the control group will be given a task that reflects traditional 'mundane' tasks. Following a four-week training period, the list sorting test will be readministered to the participants to quantify the improvement in working memory. Additionally, a questionnaire will be administered to record and evaluate participant engagement levels during the training.
Based on the preliminary data analysis of three days of training, it was observed that the reaction times either remained constant or decreased, which suggests a moderate to high level of engagement among the participants. Additionally, the accuracy consistency throughout all genres indicated that the variability between genres was not a nuisance–furthermore, overall accuracy improved by the end of the training period.
Our preliminary data suggests a positive correlation between increasing engagement in N-back memory tasks and overall working memory capability. In conclusion, the implementation of multidimensionality to enhance engagement in our modern N-back task has resulted in improved working memory among adults. Therefore, it can be inferred that the use of multidimensionality is an effective strategy to boost engagement and enhance working memory.		7:007
74
SarvagyaGoyalDowning/NjooA Comparative Analysis of Recent Progress in Blind Docking Softwares for Protein-Ligand InteractionsIn this study, we consider the advantage of blind docking as an unbiased way to identify and predict protein structures, binding sites, and their receptors to conduct a comparison of these softwares. With a thorough comparison of various blind-docking software, we aim to identify the most efficient and accurate methodology for blind-docking. We determined the scope of the analysis and comparison as including accuracy, predictive capability, algorithmic methodologies used, and the diversity of the data. Regarding the examination of these factors in an assortment of blind-docking software, each application generates information about the docking of the ligand. Thus, we analyze variables with strong correlation to the protein structures, such as the scoring function with the affinity of the ligand to the binding pocket and the estimated fitness (kcal/mol). The intention of our study is to provide a comprehensive analysis and comparison of blind-docking software for use in research pertaining to drug discovery and protein structure identification.7:505
75
RileyGuDowningDiscovering mini black holes through observations of their stellar companion in binary systemsIn April 2021, Jayasinghe et al. reported the discovery of a mass-gap black hole binary companion to the red giant V723 Mon. Building upon their astronomical models, we developed an adaptable program to uncover potential star and mass-gap black hole binary systems in large-scale datasets. Using data from ESA’s Gaia DR2 and NASA’s exoplanet archive, we employed computational analysis methods to review distortions in the stellar companion’s envelope and shifts in its radial velocity by considering the star’s increasing error in radius measurements and deviations in mean radial velocity, respectively. This resulted in the discovery of 19 potential black hole candidates at a distance of 19.5-1,100 parsecs from Earth. To create a radial velocity time series for each celestial object, we imported datasets from Carmenes, HARPS-N, among other sources. Using R’s libraries (astrochron, ggplot, and dplyr), we were able to graphically represent each star’s radial velocity time series, extracting orbital period. Referencing a mass function from Jayasinghe et al., we then approximated the mass of the potential candidates, providing strong indications that these stars exist in a binary system with a mass-gap black hole of ~3-10☉.7:202
76
SaanviGudisayJahanikiaInvestigating the Impact of Meditation on Neural Processes: Insights from EEG and Sleep AnalysisThe popularity in the research of the impact of meditation on brain activity has spiked over the past few years. Even so, analysis on the effects of mediation through electroencephalography (EEG) has been done for decades but its impacts are still uncertain. This is due to how various meditation practices affect brain activity differently, shown evident in our dataset. The dataset included data from four blocks of meditation: two thinking blocks, one breathing block, and one tradition specific meditation block. The first meditation set is a breath count meditation which may contain sleep data. The first thinking task will work as the control data to compare with the breathing task. After sorting through 50 subjects, and running it through our software titled EEGLab, the data had to be pre-processed to make accurate conclusions. With continuing data analysis through EEGLab, and identifying sleep pattern in our data using various filtering and computing methods, we hope to find the direct effects of meditation and how it may benefit cognition, perception, and emotional processing.7:507
77
NikhilGummadidalaDowningSoft Robotic Gripper for Ocean Trash CleanupWith microplastics and millions of tons of ocean trash entering the ocean each year, marine life has experienced significant repercussions. Plastics and trash that find their way into the oceans kill thousands of animals every year from entanglement and create ocean “dead zones” in which animals cannot survive. Significant progress has to be made to clean the oceans of floating debris to prevent marine life from dying. However, traditional equipment such as nets cannot efficiently collect small pieces of plastic; the use of soft robotics can mitigate this problem. Soft robots can be deployed into the ocean to pick up underwater debris that is out of the range of nets. The proposed mechanisms are a cable-driven gripper modeled on a feather star, a mothership paired with several smaller drones, and a starfish-inspired gripper. The feather star, which will be coated in adhesive, increases the probability of collecting floating trash. The mothership paired with drones is also a viable option due to its ability to cover larger areas through passive motion and pick up a variety of garbage. Each drone will have a unique gripper, making it an efficient solution to the problem at hand. The starfish-inspired gripper is cable-actuated and curves inwards to grip any larger pieces of floating debris, which helps when specific pieces of trash need to be targeted to be picked up.7:405
78
AkashGuntamaduguMcMahanVerifying the Stability of Computationally Constructed Candidate Organic MoleculesThe evaluation of molecular energies requires an understanding of wavefunctions of individual electrons in a molecule. However, due to limitations in Schrodinger’s equation when extending to multiple-electron systems, self-consistent field methods (SCFs) are used to approximate the wavefunctions and calculate energies of larger molecules. The Hartree Fock SCF is used for the computation of different molecular groups like alkanes, alkenes, and alkynes. Furthermore, isomers are evaluated with Hartree Fock, but irregularities in small cyclic molecules lead to other density functional theories used for evaluating molecular energies. Alongside research in SCF methods, a UI is created for researchers to utilize the PySCF library easily. We created a user-friendly website where researchers can upload molecular geometries to view basic details of the uploaded molecule and render the molecule in 3D. In addition, after computing a list of molecules, researchers are automatically presented with potential trends such as calculated binding energies versus number of atoms.7:302
79
AditriGuptaJahanikiaDietary patterns during COVID-19: Analyzing changes in consumption habits in adults before, during, and after COVID-19 lockdown"The COVID-19 pandemic and the resulting quarantine have caused drastic changes in an individual’s lifestyle. Dietary lifestyle, relating to an individual’s relationship with both food and exercise, is a major part of this change. We hypothesize that the pandemic caused participants to change their consumption patterns drastically, and that these changes have remained in place even after the lockdown period ended. Our questionnaire, which borrows ideas from four standardized eating assessments (SCOFF, EAT-26, CET, and EDE-Q) will aim to test this in adult participants. After completing data collection, we will conduct data analysis using R and various statistical softwares. The results of this study will provide a better understanding of one of the several changes that the pandemic has brought upon us."7:408
80
AmishGuptaDowningApplying Supervised and Unsupervised Machine Learning Techniques to Characterize Habitability in ExoplanetsFor decades, researchers have searched the night sky for habitable exoplanets. The main criterion for an exoplanet’s habitability involves its ability to host liquid water, which depends on a multitude of characteristics unique to each exoplanet and its host star, all of which must work together to create an environment capable of sustaining life. We first deduplicated the open-source NASA Exoplanet Archive dataset for efficient calculations, creating one cumulative entry for every exoplanet. Then, we calculated each host star’s bolometric luminosity and determined whether each exoplanet resided within the bounds of their host stars’ circumstellar habitable zone (CHZ). This involved attributes such as pl_orbsmax, pl_orbeccen, and st_lum. Another method in the pursuit of a supervised learning approach involved the examination of high-value candidates [e.g.: K2 18 b & K2 3 d] from prior research in regards to additional explanatory variables that might refine our algorithm's ruleset. Finally, an unsupervised model was also used to find similarities between planets and to cluster them by attribute. The data were first de-duplicated such that every measurement of a planet was averaged, after which the remaining data was run through the algorithm and represented in a 3D map. For confirmation, we compared the results of these three methods. The CHZ algorithm, in particular, corroborated two candidates with the NASA Planetary Habitability Laboratory (PHL): Kepler-62 f and Kepler-1229 b, which exhibited significant values for major explanatory variables.7:102
81
AnishGuptaJahanikiaAI Assist: Exploring the Feasibility of Biofeedback Mechanisms for Managing Symptoms of Attention-Deficit/Hyperactivity DisorderAttention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder characterized by symptoms of inattention, hyperactivity, and impulsivity that can significantly impair academic, social, and occupational functioning. Traditional interventions for ADHD often rely on student-teacher-parent interactions to manage symptoms, but the implementation and effectiveness of these interventions can vary greatly. In recent years, there has been increasing interest in exploring the use biofeedback mechanisms with Natural Language Processing (NLP)-based interventions for managing symptoms of ADHD. This presentation aims to explore the efficacy and feasibility of biofeedback mechanisms for monitoring and managing symptoms of ADHD through the use of wearable sensors, with a focus on the importance of student-teacher-parent interactions in promoting positive outcomes. By examining the current research and identifying potential areas for future investigation, this presentation seeks to provide a comprehensive overview of the potential benefits and limitations of biofeedback mechanisms for managing symptoms of ADHD, and to highlight the importance of collaboration between students, teachers, and parents in supporting individuals with ADHD in order to improve their learning experience.7:107
82
ArjunGuptaRengananthanComparing the potency of different Polyphenols in the Suppression of Amyloid-Beta Proteins in Alzheimer's disease Using Caenorhabditis Elegans as model organismAlzheimer’s disease (AD), a neurodegenerative disease that leads to cognitive memory decline, is caused by the aggregation of Amyloid-β proteins through the formation of β and γ-secretases. These secretases cut down the Amyloid-β precursor protein responsible for neural growth and repair, and result in abnormal levels of Amyloid-β (Aβ42) proteins. Natural polyphenols have antioxidative and neuroprotective properties that limit the aggregation of Aβ42 and could be used as a treatment for AD. In this study, Caenorhabditis Elegans (C. Elegans) was used as a model organism for AD as it possesses various human homologous genes and protein networks involved in the disease. The in-vivo paralysis assay evaluated the response of the nematodes in an environment containing polyphenols. The lifespan assay, which is currently in progress, will analyze worm survivability in different polyphenolic environments. From these studies, we can evaluate polyphenols with the potential to limit aggregation of Aβ42 proteins.7:402
83
ManshaGuptaAmadiDevelopment of synthetic aptamers for use as low-cost PLD1 inhibitorsPLD1 is a gene in the genome of the cell that codes for an enzyme that breaks down phosphatidylcholine into phosphatidic acid and choline. The phosphatidic acid then leaves the cell and attaches to an mTOR receptor in another cell and causes a signal to be sent into the cell. As a result of the signal transduction and a phosphorylation cascade, the cell starts performing mitosis. Due to excessive amounts of phosphatidic acid the mTOR receptor constantly signals to the cell to start mitosis through DAG kinases. PLD1 amplifies anti-apoptotic functions that cause inflammation and create a roadblock for chemotherapy. The focus of our project is to create synthetic aptamers using specialized DNA or RNA bases to act as competitive inhibitors for PLD1 allowing patients to continue chemotherapy.7:101
84
SakashGuptaMcMahanExploring Potential Solutions To The Small-Scale Problems of the Lambda Cold Dark Matter ModelThe Lambda Cold Dark Matter (λCDM), which depicts dark matter particles as cold and collisionless, is successful for large scale structures but has faced challenges on the small scale due to contradictions with astronomical observations. The challenges faced are called the small-scale problems, which are: the core-cusp problem, the missing satellites problem, and the too-big-to-fail problem. Alternate models, such as the Self Interacting Dark Matter (SIDM) model or the implementation of baryonic feedback effects, have been proposed to combat these small scale problems. Unlike λCDM particles that primarily interact through gravity, SIDM particles do interact with each other through means of self-scattering processes. In this paper, we utilize recent simulations to explore the possibility of the SIDM model with baryonic feedback combatting the small-scale problems while accurately representing the universe on the large scale to ultimately replace the λCDM model.7:204
85
SrimayiGuptaCunhaA programming package that uses current computational algorithms to predict S/MARS motifsUsing R, machine learning, and heuristics to develop a highly accurate prediction algorithm to identify S/MARS motifs. S/MARS contribute to chromatin organization and influence various metastatic processes; therefore, they are involved in colorectal cancer.7:207
86
SeoyeonHongNjooReactivity-informed Pharmacophore Editing and Biological Evaluation of Andrographolide and its Analogs7:404
87
SeoyeonHongNjooSynthesis of Dexamethasone and Related Fluorinated Corticosteroid ProdrugsSince the initial success of fluorinated corticosteroids in the 1950s, several, including dexamethasone, triamcinolone acetonide, and betamethasone, have been investigated for their potent anti-inflammatory activities and improved pharmacokinetic profiles. Among these, dexamethasone received FDA approval in 1958 and has been prescribed to millions for the treatment of inflammation, arthritis, asthma, and multiple sclerosis. Prednisolone, a similar corticosteroid, is delivered as its oxidized prodrug, prednisone, to improve its metabolic properties with a longer half-life. Inspired by this, we attempted a similar synthetic strategy on dexamethasone. We hypothesized the oxidation of its secondary alcohol at C-11 into a ketone might be an effective prodrugging strategy, as 11β-Hydroxysteroid dehydrogenase will dehydrogenate the ketone back into an alcohol. In our work, we investigate various analogs comparing dexamethasone to other fluorinated corticosteroids to evaluate the quality of prodrugging work.7:108
88
NeilHuSubramanianUsing AI to Predict Stock TrendsWe are creating an AI model that can predict stock trends and fluctuations. We want to start by creating a model and feeding it data on Netflix's stock, and then give it data from other stocks so that eventually it can predict the future stock market. We can accomplish this by using softwares such as ARIMA and Prophet, and testing using RMSE.7:303
89
DanielHuangJahanikiaAI Assist: Exploring the Feasibility of Biofeedback Mechanisms for Managing Symptoms of Attention-Deficit/Hyperactivity DisorderAttention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder characterized by symptoms of inattention, hyperactivity, and impulsivity that can significantly impair academic, social, and occupational functioning. Traditional interventions for ADHD often rely on student-teacher-parent interactions to manage symptoms, but the implementation and effectiveness of these interventions can vary greatly. In recent years, there has been increasing interest in exploring the use biofeedback mechanisms with Natural Language Processing (NLP)-based interventions for managing symptoms of ADHD. This presentation aims to explore the efficacy and feasibility of biofeedback mechanisms for monitoring and managing symptoms of ADHD through the use of wearable sensors, with a focus on the importance of student-teacher-parent interactions in promoting positive outcomes. By examining the current research and identifying potential areas for future investigation, this presentation seeks to provide a comprehensive overview of the potential benefits and limitations of biofeedback mechanisms for managing symptoms of ADHD, and to highlight the importance of collaboration between students, teachers, and parents in supporting individuals with ADHD in order to improve their learning experience.7:107
90
JustinHuangDowning/NjooA Comparative Analysis of Recent Progress in Blind Docking Softwares for Protein-Ligand InteractionsIn this study, we consider the advantage of blind docking as an unbiased way to identify and predict protein structures, binding sites, and their receptors to conduct a comparison of these softwares. With a thorough comparison of various blind-docking software, we aim to identify the most efficient and accurate methodology for blind-docking. We determined the scope of the analysis and comparison as including accuracy, predictive capability, algorithmic methodologies used, and the diversity of the data. Regarding the examination of these factors in an assortment of blind-docking software, each application generates information about the docking of the ligand. Thus, we analyze variables with strong correlation to the protein structures, such as the scoring function with the affinity of the ligand to the binding pocket and the estimated fitness (kcal/mol). The intention of our study is to provide a comprehensive analysis and comparison of blind-docking software for use in research pertaining to drug discovery and protein structure identification.7:505
91
MelinaHuangMcMahanAdvancing Environmental Mapping and Forest Health Assessments: Integrating Machine Learning Algorithms in Autonomous DronesWith the increasing concern for environmental conservation, there is a growing need for efficient methods of environmental mapping and forest health assessments. However, traditional methods employed by the U.S Forest Health Monitoring have faced controversy due to limited spatial resolution and integration of modern technologies. This research paper explores the application of machine learning algorithms in autonomous drones to conduct forest health assessments. Autonomous drones have the ability to collect timely, up-to-date data, which offers enhanced accuracy. This study focuses on training Deep Learning (DL) models to classify different environmental features based on aerial imagery captured by drones. To achieve accurate and efficient data collection, we will utilize Red-Green-Blue imaging and Convolutional Neural Networks (CNN) with the appropriate evaluation metrics, such as the Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI), and foliage color, to create tree classes and identify forest health indicators. By integrating machine learning algorithms into forest health assessment, this study provides a more efficient, accurate, and up-to-date approach to monitor and evaluate the well-being of forests—supporting ongoing efforts towards environmental management and conservation.7:003
92
AdarshIyengarBrahDocking Ligands using SwissDock and Autodock to learn about the most mutable ligands in PI3KThe purpose of this experiment is to find out which ligands are the most easily mutable which we can tell by checking the ligand binding affinities using Swissdock and Autodock. Of which Maestro was used as a modelling software to analyze physical characteristics to look at for future ligands to test.8:001
93
DiyaJainMcMahanGenerating Chemically Stable Molecules via Quantum Computing and Python Molecular Benchmarking ProcessesCurrent drug discovery and development processes can cost upwards to billions of dollars and last five to twelve years to create one FDA-approved drug, so researchers have been implementing computational chemistry methods to current molecular synthesis pathways. However, current computational chemistry methods often have inefficient runtimes, as the chemical space is vast. Aiming for a more efficient runtime and robust analysis of high-dimensional molecular data, our group previously implemented the Hybrid Quantum-Classical Generative Adversarial Network (QGAN) and the most recent iteration of the model is the Hybrid Quantum-Classical Graph Generative Adversarial Network (QNetGAN) to synthesize chemically feasible molecules. The QNetGAN addresses the issues faced by the QGAN including the distance between atoms exceeding bonding length, which causes most of the atoms in the generated molecules to remain unbonded. By generating graphs and utilizing long-short term memory cells, QNetGAN generated 141/300 structurally valid molecules that satisfy Lipinski’s Rule of Five, yielding a 47% success rate — a notable increase from the prior 2.3% — with a minimal training time of 10.164 minutes. However, molecules generated by QNetGAN remain a work in progress, with the majority of molecules failing to satisfy the Octet Rule and having unoptimized bond lengths and bond angles. Our work focuses on implementing chemical post-processing algorithms to increase the chemical structures’ stability and feasibility. Our Python-based Octet Rule algorithm employs a Depth-First Search (DFS) traversal structure to count the number of bonds and open orbitals on each atom. Then, our formal charge calculation algorithm calculates the formal charge of each atom, to find the best overall structure of the molecule with the most stability. Finally, our Hydrogen Addition Algorithm builds upon the Octet Rule algorithm to traverse through the molecule’s adjacency matrix, adding hydrogens to central atoms where necessary to complete the molecule. Although a work in progress, our post-processing algorithms have been able to check for Octet Rule satisfaction, calculate formal charge, and add hydrogen atoms to molecules as necessary with 100% accuracy. In the future, we plan to continue testing our algorithms on larger molecular structures, while implementing methods to handle exceptional cases to the basic rules. Furthermore, as our generated molecules are initially only given 2D coordinates due to software limitations, we are working to recalculate each atom’s coordinates such that the final molecular structure corresponds correctly to their molecular geometries (i.e. tetrahedral, trigonal planar, bent, etc.) and bond angles.7:103
94
RiniJainJahanikiaPersonalityGPT: Closing Communication Gaps in Texting using Artificial IntelligenceIn the realm of AI-driven communication, PersonalityGPT is a groundbreaking solution that leveraging the influential Big Five personality traits to bridge communication gaps across various personality types. This innovative model integrates with users' emotional states, aided by sentient AI components, enhancing empathy and understanding. By seamlessly adapting its responses to different personality types, the model provides more authentic and meaningful conversations. Furthermore, the integration of sentient AI components amplifies its empathetic capabilities, enabling it to evaluate and respond to subtle emotional states, thereby fostering a deeper understanding of users' mental health, a field that is becoming ever more important.7:205
95
RishabJainLiuAn Explainable AI (XAI) Model that Detects Pneumonia From X-Ray ImagesOur project centers on utilizing Explainable AI (XAI) methods on our CNN models to detect pneumonia in X-ray lung images. Leveraging techniques such as LIME, Saliency Maps, Grad-CAM, Occlusion Tests, Deep Taylor Decomposition, and SHAP, we were able to visualize the results of our models. Our XAI-focused approach enhances transparency, enabling insights into the neural network decision process. By combining multiple XAI methods to X-ray images, we successfully distinguished between pneumonia-affected and healthy lung conditions.7:401
96
ShriyaJainNjooReactivity-informed Pharmacophore Editing and Biological Evaluation of Andrographolide and its AnalogsAndrographolide, a natural product labdane diterpenoid extracted from the plant Andrographis paniculata, is known to have potent anti-cancer activity. The putative mode of action of andrographolide is the inhibition of Nf-kB, which subsequently leads to downregulation of a myriad of cell signaling pathways typically involved in cell cycle regulation. However, previous research has suggested that chemical modifications to the C19 hydroxyl and C17 alkene may alter the biological target of the andrographolide analog to the Wnt/𝜷-catenin signaling pathway. With this pharmacophore in mind, we designed a library of targeted andrographolide C19 and C17 analogs with altered polarity and steric profiles to probe the effects of large, hydrophobic silyl and trityl ethers at C19 and epoxidations of C17 on metabolic stability and the primary mechanism of action. After studying the potency of our analogs through MTT and cell migration assays, we assessed the analogs’ downstream transcriptomic effects on key apoptosis-regulating pathways and their potential as a protein inhibitor in the Wnt/𝜷-catenin signaling pathways. Additionally, we installed a number of benzyl acetals to the a-ring of the andrographolide scaffold. These functionalities, being hydrolyzable under mildly acidic conditions, led us to hypothesize that they would function as hydrolytically labile prodrugs of andrographolide. In order to study the reaction dynamics of this process, we synthesized a library of 4-substituted benzaldehyde acetals and present a hammett linear free energy relationship of these prodrugs. We determined that the 4-chloro is the most hydrolytically stable.7:404
97
RitwikJayaramanJahanikiaDecoding Inner Speech with Brain-Computer Interfaces: A Study in EEG Analysis and Machine LearningBrain-Computer Interfaces (BCIs) identify brain signals and convert them into instructions executed by external devices. BCIs have the potential to significantly enhance the quality of life for individuals afflicted with neuromuscular disorders by restoring lost functionality. These disorders often hinder communication abilities, necessitating BCIs capable of deciphering internal speech. Electroencephalography (EEG) is a commonly used noninvasive neuroimaging technique that measures the brain's electrophysiological responses resulting from synchronized neurons. Recent advancements in machine learning have facilitated the detection of brain patterns in EEG data, leading to more promising and dependable BCIs. In this project, we employ a dataset comprising EEG data containing inner speech commands from 10 participants. Through data analysis and the application of machine learning, our objective is to create a model that can accurately interpret inner speech.7:106
98
DaliaJazrawiJahanikiaInvestigating the Impact of Meditation on Neural Processes: Insights from EEG and Sleep AnalysisThe popularity in the research of the impact of meditation on brain activity has spiked over the past few years. Even so, analysis on the effects of mediation through electroencephalography (EEG) has been done for decades but its impacts are still uncertain. This is due to how various meditation practices affect brain activity differently, shown evident in our dataset. The dataset included data from four blocks of meditation: two thinking blocks, one breathing block, and one tradition specific meditation block. The first meditation set is a breath count meditation which may contain sleep data. The first thinking task will work as the control data to compare with the breathing task. After sorting through 50 subjects, and running it through our software titled EEGLab, the data had to be pre-processed to make accurate conclusions. With continuing data analysis through EEGLab, and identifying sleep pattern in our data using various filtering and computing methods, we hope to find the direct effects of meditation and how it may benefit cognition, perception, and emotional processing.7:507
99
ElianeJuangDowning/NjooA Comparative Analysis of Recent Progress in Blind Docking Softwares for Protein-Ligand InteractionsIn this study, we consider the advantage of blind docking as an unbiased way to identify and predict protein structures, binding sites, and their receptors to conduct a comparison of these softwares. With a thorough comparison of various blind-docking software, we aim to identify the most efficient and accurate methodology for blind-docking. We determined the scope of the analysis and comparison as including accuracy, predictive capability, algorithmic methodologies used, and the diversity of the data. Regarding the examination of these factors in an assortment of blind-docking software, each application generates information about the docking of the ligand. Thus, we analyze variables with strong correlation to the protein structures, such as the scoring function with the affinity of the ligand to the binding pocket and the estimated fitness (kcal/mol). The intention of our study is to provide a comprehensive analysis and comparison of blind-docking software for use in research pertaining to drug discovery and protein structure identification.7:505
100
AnishJupudyCunhaComparative Genomic Analysis of Colorectal Cancer Microbiome Bacteria to Discover Novel RelationshipsColorectal cancer (CRC) is uncontrolled tumor growth that starts in the rectum or colon (Park E. et al., 2022). Many factors affect the development of cancer, including daily habits, environments, and genetics. Our research focuses on analyzing the differences in pathways/enzymes between cancerous and non-cancerous associated bacteria in the gut microbiome outlined by a recent cancer microbiome review (Park E. et al., 2022). By utilizing the Bacterial and Viral Bioinformatics Resource Center (BV-BRC), we compiled our bacteria’s genetic information into genome groups and used the comparative systems service to identify target pathways and construct phylogenetic trees. After focusing on genomes, we delved deeper into the enzymes. The programming language R was used to narrow down four specific enzymes from the set of genomes: two from the pathways only in non-cancerous bacteria and two in cancerous-associated bacteria. A Multiple Sequence Alignment (MSA) run at the genome level identified the range of lowest entropy among the genes in the four enzymes - one of which had the lowest range of 30-40. We are using NCBI Blast and other bioinformatics methods to characterize/validate the four enzymes in our respective target bacteria. Our end goal is to target/screen the unique pathways and enzymes (like the enzyme with EC number 5.4.3.2) of the cancer-associated bacteria and non-cancerous associated bacteria to decrease the metastasis of CRC tumors (Park E. et al., 2022). These genes, that help create the enzymes, can be manipulated in the wet lab as shown by the cited paper.7:307