2015 MU theses - Access restricted to UM
Permanent URI for this collection
The items in this collection are theses that are available only to members of the University of Missouri system. Click on one of the browse buttons above for a complete listing of the works.
Browse
Recent Submissions
Item Linking wetland management decisions to secretive marsh bird habitat use during spring migration and summer breeding on public wetlands in Missouri(University of Missouri--Columbia, 2015) Hill, Evan B.; Webb, Elizabeth B.[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] The extent to which secretive marsh birds rely on wetlands in Missouri during spring migration is relatively unknown. My objective for chapter one was to determine how initial marsh bird occupancy and subsequent colonization and departure probabilities were influenced by wetland management practices, including the duration and initiation date of spring water-level drawdowns, and associated wetland habitat characteristics. We used dynamic occupancy modeling to evaluate factors that influence SMB occupancy and colonization/departure probabilities. Sora and American bittern occupancy models indicated a positive relationship between occupancy probability and duration of drawdown, however the top occupancy model for Virginia rail was the null model. The top colonization/departure model for sora included vegetation density and percent of a site containing emergent vegetation, with both variables having a positive relationship with colonization probability and a negative relationship with departure probability. The top colonization/departure model for Virginia rail included range of water depth and range of vegetation height, both of which had a negative relationship with colonization and departure probability. The top colonization/departure model for American bittern included vegetation interspersion, whereas the top model for least bittern included the percent site inundated and overall area inundated. My objective for chapter two was to determine effects of hydrologic management and habitat characteristics on habitat selection and the daily survival rate (DSR) of least bittern on public wetlands in Missouri at two scales: the entire wetland and the nest point. Least bittern populations have been in decline since the 1970s, most likely due to extensive loss of freshwater emergent wetlands, the primary nesting habitat of least bittern. The decline in nesting habitat emphasizes the need for effective wetland management within the nesting range of least bitterns. The extent to which least bittern rely on wetlands in Missouri during summer nesting efforts is poorly understood. The logistic exposure method was used to evaluate DSR as a function of covariates. At the wetland scale, logistic regression was used to evaluate models composed of combinations of covariates thought to influence least bittern nest site selection. The percent of a wetland covered in emergent vegetation and the average water depth were positively associated with probability of selection at the wetland scale. At the point scale, discrete choice was used to evaluate models composed of combinations of covariates thought to influence least bittern nest site selection. The relative probability of use was positively related with water depth, percent of a site in emergent vegetation, and negatively related with vegetation density. Daily Survival Rate was positively related with average water depth at nest points. These results are important to inform management decisions intended to create wetland conditions favorable to SMBs. Both migrants and breeders are more likely to use wetlands with emergent vegetation interspersed with patches of open water. Drawdown schedules will increase occupancy if they are timed to conform to the life history stage of the target species, providing water for migrants in April-May, and for breeders in June and July. A possible ecological trap will be avoided if drawdowns are complete before least bittern begin nest site selection in mid to late May.Item The association of externalizing behavior with individual DSM-5 alcohol use disorder criteria(University of Missouri--Columbia, 2015) McDowell, Yoanna E.; Sher, Kenneth J.[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] The externalizing spectrum contains a range of disinhibition-related conditions. Comorbidity among externalizing disorders is commonly investigated at the syndromal and trait level precluding insight into the relationship of symptoms across externalizing disorders. It is unknown whether comorbidity across externalizing disorders holds constant across highly varied, individual AUD criteria ranging from symptoms reflecting neuroadaptation (e.g., tolerance) to symptoms reflecting behavioral changes (e.g., social problems). The present study aimed to determine the degree to which DSM-5 alcohol use disorder (AUD) criteria are associated with symptom endorsement from two externalizing disorders. Psychometric inquiries via multivariate and factor analytic approaches estimated relative and unique associations of externalizing behavior on AUD criteria endorsement. Our results indicate modest relations of externalizing behavior and AUD criteria endorsement. For example, social problems and role interference criteria were most strongly associated with externalizing behavior across analytic approaches, with general and unique associations between externalizing behavior and social problems. Additionally, tolerance was most weakly associated with externalizing behavior across approaches. Results highlight potential etiological heterogeneity among AUD criteria that could guide future diagnostic refinements and treatment methods.Item MEMS parallel plate variable capacitor for energy harvesting(University of Missouri--Columbia, 2015) Chang, Yushan; Almasri, Mahmoud[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] In recent years, there is a dramatic increase in use of portable, wireless sensors and actuators which are widely powered by batteries. However, these batteries have drawbacks such as limited energy, requiring replacement, maintenance, and recharges which could cause extra manpower, time and financial expenses. In order to address this problem, a novel MEMS parallel plate variable capacitor for energy harvesting was developed to extract (harvest) wasted energy from undesired mechanical vibrations. This makes it possible to have devices perpetual self-powered. The design, fabrication and characterization of this MEMS parallel plate variable capacitor is presented in details in this thesis. The result shows a promising future of this energy harvester.Item A framework for histopathology image segmentation and classification(University of Missouri--Columbia, 2015) Al-Milaji, Zahraa; Ersoy, Filiz Bunyak[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] It is increasingly appreciated that tumor-stroma plays an integral part in cancer initiation, growth, and progression. Recently, it has been shown that the stromal elements of tumors hold prognostic as well as response-predictive information. Automated analysis of histopathology images have a great potential for both clinical applications (e.g., to reduce/eliminate inter- and intra-observer variations in diagnosis) and research applications (e.g., to understand the biological mechanisms of the disease process). This thesis proposed two computational image analysis frameworks to identify epithelial versus stromal tissue regions in images of Hematoxylin and Eosin (H and E) stained breast cancer specimens. The first framework used handcrafted image features with a support vector machine classifier. The H and E stained images were first segmented into coherent partitions/superpixels; then a number of regional color and texture features were extracted from these partitions; and finally a support vector machine classifier was trained to classify the image regions into epithelium and stroma classes. The second framework introduced a novel feature extraction and machine learning approach for identification of epithelium and stroma regions using deep learning methods. Deep convolutional neural networks (CNNs) were trained to extract hierarchical features from raw pixels of H and E stain images and to perform epithelium versus stroma classification. The classification results from deep convolutional neural networks were further fused with image segmentation results for improved performance. The proposed frameworks were evaluated on images from Stanford Tissue Microarray database. Deep learning based framework produced comparable results to the classification framework that used carefully hand-crafted image features. Experimental results showed that incorporation of an explicit segmentation step increased the classification accuracy in both frameworks.Item Complexity reduction and visualization tool for RDF knowledge network application in precision medicine(University of Missouri--Columbia, 2015) Al-Taie, Zainab; Xu, Dong, 1965-; Joshi, Trupti[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] Precision medicine is one of the most rapidly emerging areas of research and development, crucial for improving patient care, but there is a lack of a comprehensive set of tools that are easy to use for analysis and incorporating genomic data into clinical decision. In our study, we attempt to reconstruct interrelationships among biomarker proteins, diseases and signal transduction pathways for individualized treatments. Towards this, we have developed a suite of tools, which can shed some light on tumorigenesis and genomic changes taking place in individual patient. Firstly, we have developed a visualization and curation tool to resolve the problem of missing information in the KEGG database. This tool helped in curation of KEGG pathways. The curated pathways have been converted into RDF to create a knowledge base network. Secondly, we have developed Complexity Reduction and Visualization (CRV) tool for pathologists, oncologists and other specialists in precision medicine. This tool reduces the complexity of the knowledge network by finding the shortest paths among a set of start genes and end genes. The resulting network will be visualized using d3.js. Such a suite of tools can be applied to answering diverse questions including getting a better understanding of genomics mechanisms that play a role in metastasis vs. nonmetastatic cancers. We have applied this to Lung Adenocarcinoma (LUAD) samples. Our system helps build hypothesis but needs to be further validated with more testing using some benchmark ground truth datasets. In one dataset, we have found 90 genes and 15 pathways like focal adhesion, mTOR and ErbR signaling pathway that have a role in transforming the cancer from non-metastatic cancer to be a metastatic cancer.