Electronic Thesis or Dissertation

The goal of this dissertation was to examine the fate and fragmentation of plastic debris in the marine environment and characterize the microbial communities colonizing naturally occurring substrates and geologically recent plastic inputs in the ocean using analytical chemistry and genomic techniques. Altogether, the data presented herein reveal the presence of heretofore undescribed plastic accumulation zones in the ocean and depict a stark contrast in microbial traits between early and mature plastic biofilm stages. These data further suggest that emergent plastic biofilm properties can be forecasted across environmental gradients, with the largescale genomic characteristics of early colonizers varying little across conditions. Chapter I of this thesis is an introduction to the current body of work regarding the plastisphere. Chapter II explores the ‘cradle to grave’ fragmentation, transformation, and transport of model microplastic particles and single-use plastic items in an artificial beach setting. Chapter III presents the first metagenomic insights into early biofilm formation on virgin microplastic surfaces in the marine environment and how early colonizers self-assemble, compared to mature, taxonomically, and metabolically diverse biofilms residing on free-drifting plastic pollution. Chapter IV further investigates microbial adaptations for initial colonization of virgin control and plastic surfaces and examines biofilm assemblage dynamics by employing metagenomics on a 16-day time series in a wastewater treatment facility. Chapter V synthesizes observations from the previous core chapters and discusses what these findings mean in a broader ecological and evolutionary perspective. Appendix 1 is a reprint of the manuscript describing the distribution of microplastics beneath the inner and outer plastic accumulation zones of the South Atlantic Subtropical Gyre. Appendix 2 is a submitted manuscript detailing the isolation, characterization, and selective adaptations of Vibrio bacteria colonizing eel leptocephali, free-drifting plastic pollution, Sargassum, and seawater in the North Atlantic Ocean.

Many marine species distributions have shifted poleward in response to global climate change. Many environmental characteristics will be affected by climate change including temperature and phytoplankton concentration; yet, photoperiod will remain the same. It is imperative to gather baseline distribution data on migratory species so that these shifts can be measured and mitigated. Sex-specific differences in reproductive strategies cause sexual segregation, sex-specific differences in spatial distribution. Female blacktip sharks exhibit a synchronous, biennial reproductive cycle in which one year of reproduction is followed by a resting year. Acoustic telemetry can be used in conjunction with collaborative networks to track migratory species over great distances. However, the irregular spacing of acoustic receivers often results in sporadic detection data, which can lead to skewed distribution information. This project developed and tested an analysis process to regularize sporadic acoustic detection data. Those regularized data were then applied to cluster analyses to determine the seasonal spatial distributions of blacktip sharks, Carcharhinus limbatus, off the United States East Coast and corresponding environmental correlates of latitudinal movement. Sexes of this population were investigated separately and in combination. Differences in distribution were evaluated between sexes, and within females, between reproductive states. These data showed that the U.S. East Coast blacktip shark population distributes from Palm Beach County, FL to Long Island, NY and exhibits sexual segregation, in which females display a more truncated migratory pattern than males.

The desire to own human skeletal remains has been prevalent for many years; in our modern technological age avenues for this market have exploded across the internet. This research focuses on Facebook groups dedicated to oddity sales and collecting. Purchasing human remains is illegal in Georgia, Louisiana, and Tennessee as well as prohibited by Facebook terms of service, but these sales persist.
Over the course of 2021, 319 listings for human skeletal remains were recorded across six Facebook groups. These listings accounted for most skeletal elements found within the human skeleton. Many elements are artistic in nature, something viewed as “Giving a second life” to the remains, as observed within these groups. To fully understand the driving force behind this market requires cultural insight about the perception of human remains as well as the culture found within these groups. Kinship, friendship, and trust are all clearly expressed between buyers and sellers.

The lack of physiologically relevant human esophageal cancer models has as a result that many esophageal cancer studies are encountering major bottleneck challenges in achieving breakthrough progress. To address the issue, here a 3D esophageal tumor tissue model was engineered using a biomimetic decellularized esophageal matrix in a customized bioreactor. To obtain a biomimetic esophageal matrix, a detergent-free, rapid decellularization method was developed to decellularize porcine esophagus. The decellularized esophageal matrix (DEM) was characterized and the DEM was utilized for the growth of esophageal cancer cell KYSE30 in well plates and the bioreactor. Then the expression of cancerrelated markers of KYSE30 cells was analyzed and compared with formalin-fixed, paraffin-embedded (FFPE) esophageal squamous cell carcinoma (ESCC) tissue biospecimens. Results show that the detergent-free decellularization method preserved the esophageal matrix components and effectively removed cell nucleus. KYSE30 cancer cells proliferated well on and inside the DEM. KYSE30 cells cultured on the DEM in the dynamic bioreactor show different cancer marker expressions than those in the static well plate, and also share some similarities to the FFPE-ESCC biospecimens.

This thesis outlines the cultural and literary interpretations of Haitian folktales among Haitians, Haitian Americans, and Americans of non-Haitian descent. This thesis seeks to explain how folktales can be used to highlight cultural identity through symbolic analysis, cultural reflexive theory and a cross-cultural analysis model. The nuanced differences found in the reception of the folktales that are associated with the characters of Uncle Bouki and Ti Malis by the three research groups forms the basis of this thesis research design. The characters of Uncle Bouki and Ti Malis are, in effect, cultural literary examples of how folktales could be used to explain Haitian rural societal values or norms

This thesis presents an analysis of Steven Spielberg’s Jaws and the film’s depiction of nature. This analysis will show that the film derives horror from the depiction of nature encroaching on human spaces. Through the film’s depictions of shark attacks, it forces viewers to confront their own edibility. The filmmaking techniques place humans on the other side of the eater/eaten binary, and present humans a prey. Similarly, the depictions of environments show the presence of nature as a disruption to the film’s established visual style. This thesis asserts film analysis as a necessary tool in understanding the nature/culture binary and how film narratives can contribute to this division.

Cyber attack is a strong threat to the digital world. So, it’s very essential to keep the network safe. Network Intrusion Detection system is the system to address this problem. Network Intrusion Detection system functions like a firewall, and monitors incoming and outgoing traffic like ingress and egress filtering fire wall. Network Intrusion Detection System does anomaly and hybrid detection for detecting known and unknown attacks. My thesis discusses about the several network cyber attacks we face nowadays and I created several Deep learning models to detect accurately, I used NSL-KDD dataset which is a popular dataset, that contains several network attacks. After experimenting with different deep learning models I found some disparities in the training accuracy and validation accuracy, which is a clear indication of overfitting. To reduce the overfitting I introduced regularization and dropout in the models and experimented with different hyperparameters.

Gubernatorial mansions are historically home to white men. Candidates with contrary identities, such as women and racial minorities have met limited success in their quest for office. Yet the number of women seeking executive level office has increased and these women represent a broader intersectional identity. The low percentage of women governors has been examined in detail, but that analysis largely holds gender as an isolated variable and does not consider the candidates' broader identity. This project posits gender is only one factor of candidate identity called into question when it is nonnormative and varies from historical office holders. I argue candidate identity interacts with the identity expressed by voters and the collective identity found in social movements. The ability of the candidate to navigate this interaction and use it to their advantage is paramount to their success. I find that structural differences in the Democratic and Republican parties provide opportunities and constraints for women candidates. Further, gender, race, and previous political experience are intersectional and create different responses by candidates. Ultimately, successful candidates align their political identity with the collective identity found in contemporary social movements as a mitigation mechanism for voters uncomfortable with who the candidate appears to be.

Sexual violence (SV) is a significant problem that impacts women on college campuses at an alarming rate (Fischer et al., 2018). The body of research published regarding women’s experiences with SV on college campuses disproportionately focuses on Caucasian women (Oney, 2018). Few studies address the specific concerns of minority women and their experiences with SV and even fewer studies serve to identify contributing factors to their recovery. In addition to the prevalence of SV on college campuses, the rates of rape myth acceptance (RMA) that have been studied among this age group focus primarily on White cisgender men and women, and again, are understudied in women who identify as racial/ethnic minorities (Oney, 2018).
Research correlates high levels of RMA with a decreased willingness to accept recovery-promoting assistance post-SV, which reduces a survivor’s willingness to access to services such as counseling (Oney, 2018). The objective of this study was to determine if rape myth acceptance predicts recovery self-efficacy and if experiences of SV serve as a mediating variable between recovery-self efficacy and RMA in ethnic and racial minority college-age women.
The results of this study indicate that RMA does not predict or mediate the variables of recovery self-efficacy and SV. A linear regression analysis was used to establish if RMA predicts recovery self-efficacy, the factors within the scales were not correlated and additional tests yielded non-statistically significant results; (b = -0.02, t = -0.29, p = .77). The study also was unable to provide evidence of experiences of SV being a mediating variable between RMA and recovery self-efficacy through a mediation analysis (b =.00, SE = .002, 95% CI = [-.004, .004], p =.89).

Advancements in Artificial Intelligence (AI) and Machine Learning (ML) have significantly improved their application in dermatology. However, bias issues in AI systems can result in missed diagnoses and disparities in healthcare, especially for individuals with different skin types. This dissertation aims to investigate and improve the fairness and bias in machine learning models for dermatology by evaluating and enhancing their performance across different Fitzpatrick skin types.
The technical contributions of the dissertation include generating metadata for Fitzpatrick Skin Type using Individual Typology Angle; outlining best practices for Explainable AI (XAI) and the use of colormaps; developing and enhancing ML models through skin color transformation and extending the models to include XAI methods for better interpretation and improvement of fairness and bias; and providing a list of steps for successful application of deep learning in medical image analysis.
The research findings of this dissertation have the potential to contribute to the development of fair and unbiased AI/ML models in dermatology. This can ultimately lead to better health outcomes and reduced healthcare costs, particularly for individuals with different skin types.