Copyright (c) 2017 University of Miami All rights reserved. http://scholarlyrepository.miami.edu/oa_dissertations Recent documents in Open Access Dissertations en-us Sun, 30 Jul 2017 01:34:01 PDT 3600 http://scholarlyrepository.miami.edu/oa_dissertations/1919 http://scholarlyrepository.miami.edu/oa_dissertations/1919 Fri, 28 Jul 2017 13:54:53 PDT In my dissertation, I trace the dramatic change that Spanish society experienced around scientific and technological developments from the final decades of Franco’s dictatorship (1939-1975) until now. I analyze how the recent changes and the emerging fascination with science and technology are borne out by the works of three different clusters of writers: a) the formal innovators of the 1960s; b) the pop writers of the 1990s; and c) a group of fiction writers of the 2000s imbued by techno-science (the so-called mutantes). I combine these analyses with a study of the efforts to popularize science developed by the Francoist government; the style of popularization employed by Eduard Punset in the democratic period; and, finally, the innovative and participatory techniques of popularization developed in the Medialab Prado, and their relation to the construction of new collective cultural values in contemporary Spain. The final goal of my dissertation is to craft a map that helps to explain changes in Spain. My work indicates that there has been an epistemic change among dominant Spanish elites. My dissertation tries to answer whether this epistemic change has taken place in contemporary Spain and if so, why now and not in the past.

]]> Carlos Gamez-Perez http://scholarlyrepository.miami.edu/oa_dissertations/1918 http://scholarlyrepository.miami.edu/oa_dissertations/1918 Thu, 27 Jul 2017 08:19:02 PDT Parkinson disease (PD) is a neurodegenerative disease characterized by the loss of dopaminergic neurons of the substantia nigra. Parkinson disease related variants have been identified in many different genes, and the products of these genes are involved in autophagy, mitochondria and cell stability pathways. However, these identified variants are responsible for only 15% of PD cases, with the bulk of these cases being idiopathic PD. While genes do play a role, the most important risk factor in PD is age. Though the average age of onset for PD is 62, early onset forms of PD have been diagnosed in patients even prior to 40 years of age. This poses the question, how does genetics play a role in PD dysfunction, and is the dysfunction that causes PD present throughout life, or does it only manifest in later stages. The advent of patient-specific induced pluripotent stem cell technology (iPSCs) allows for a unique model system for disease study. Though useful in furthering our understanding of the disease, previous model systems lacked the human genetic background relevant to PD. iPSCs derived from patients with confirmed disease status, and in some cases PD-associated variants can be differentiated into cell types of interest for in-vitro characterization. Identification of commonalities between the many forms of PD could be valuable in identifying targets which will aid in the development of novel therapeutic strategies. Patient-specific iPSC lines were derived from individuals with PD. These lines were subsequently differentiated into midbrain dopaminergic neurons. In the first study, individuals carrying variants in GBA (beta-glucocerebrosidase) exhibited an early temporal phenotype. GBA-variant lines induced from neuronal aggregates consistently displayed deficits in neurite outgrowth after five days of differentiation. When compared to control lines these cells exhibited migration from the tight aggregates and shorter neurite growth from aggregates. This phenotype is less prevalent as cells were passaged from aggregate culture, and in non-aggregate based differentiation scheme this phenotype was not observed. The findings of our transcriptomic and pathway analysis indicate that patient derived PD lines showed altered expression in genes involved in extracellular matrix receptor interaction, focal adhesion, and p53 signaling, when compared to age-matched controls. Furthermore, this is consistent with transcriptomic studies utilizing post-mortem PD patient brains. In total, our data indicates that in the case of GBA-variant lines the dysfunction that leads to PD may be present early on. This could account for the fact that many GBA carriers with PD suffer from early onset PD, and in some cases, a more severe form of the disease. The transcriptomic study highlighted differential expression in genes related to focal adhesion, and extracellular matrix receptor interaction dysfunction in these gene sets has been previously demonstrated in other systems. Additionally, these pathways have been linked to neuronal cellular instability and subsequent cell death seen in PD.

]]> Kinsley Belle http://scholarlyrepository.miami.edu/oa_dissertations/1917 http://scholarlyrepository.miami.edu/oa_dissertations/1917 Thu, 27 Jul 2017 08:18:58 PDT I study different types of statistical solutions (Hopf, Foias , Vishik-Fursikov) for nonlinear evolution equations. As a test equation, I use the nonlinear Schrödinger equation with power-like nonlinearity in the case where the proofs of uniqueness are not available. When there is no uniqueness in the original equation, statistical solutions are not unique. For autonomous differential equations, there is a formal semigroup property. I propose to look for statistical solutions with an analogous property. For statistical solutions, this should be the homogeneous Markov property. I call such solutions Markov statistical solutions. The proofs of the existence of the Markov statistical solutions rely on the Markov selection theorem. N.V. Krylov was the first to realize the importance of the Markov selection in the context of Markov processes. D.W. Stroock, and S.R.S. Varadhan re-framed Krylov’s selection in the context of solutions to the martingale problem. Recently, their results have been used by F. Flandoli and M. Romito, and Goldys et al., for the analysis of the Navier-Stokes equation with additive noise. I use the Markov selection theorem to prove the existence of Markov statistical solutions. I give a new proof of the Markov selection theorem. This proof has prompted me to look back at the set-valued solutions of deterministic equations, where the analog of the homogeneous Markov property should be the semigroup property. It turned out that no theorems of existence of selections satisfying the semigroup property were known. I state and prove a selection theorem for measurable selections with the semigroup property. Such result is important in its own right. I use it here to give a second proof of the existence of Vishik-Fursikov measures.

]]> Jorge Eduardo Cardona http://scholarlyrepository.miami.edu/oa_dissertations/1916 http://scholarlyrepository.miami.edu/oa_dissertations/1916 Thu, 27 Jul 2017 08:18:53 PDT Novel algorithm design, implementation, and optimization in histone modification ChIP-seq analysis of broad chromatin mark data is the subject of part I of this dissertation, focusing on data-driven precision medicine computational strategies for mapping ChIP-seq peaks to genomic features (and biological function) as well as coverage island analysis of low-sample size ChIP-seq experiments within individual biological replicates. Part II of this dissertation focuses on novel algorithm design, implementation, and analysis of high performance visualization techniques for histone modification ChIP-seq data using static and interactive biological gene expression heatmaps.

]]> Bohdan Khomtchouk http://scholarlyrepository.miami.edu/oa_dissertations/1915 http://scholarlyrepository.miami.edu/oa_dissertations/1915 Tue, 25 Jul 2017 09:11:18 PDT Mexico’s traditional role in the illegal drug trade has primarily been as a local producer and exporter of marijuana and heroin for the U.S. market, but, since the mid 1980’s, Mexico has become an increasingly popular transshipment point for South American cocaine. Thus Drug trafficking, has had insidious effects on U.S.-Mexican relations. By and large there has been no clear trend underlying the ways in which Mexico and the United States have related to one another on the issue of the drug trade. There have been periods marked by cooperative initiatives, as there have been periods when the U.S. has dominated the relationship and in so doing determined the anti-drug efforts by both countries. While most scholars tend to study drug policy in Mexico in terms of a US policy byproduct this dissertation suggests that drug policy in Mexico is a intermestic game. Thus -and contrary to conventional wisdom- this research hypothesizes that nowadays Mexican drug policy is not only shaped by U.S. power but by convergent interests, perceptions and identities -socially constructed- by the governing elites in each country. This dissertation draws on the main theoretical approaches to International Relations and Comparative Politics, and despite the fact that it is a case study, it aims to provide a sufficiently abstract explanation to understand drug policy in different settings.

]]> Jorge Rebolledo Flores http://scholarlyrepository.miami.edu/oa_dissertations/1914 http://scholarlyrepository.miami.edu/oa_dissertations/1914 Tue, 25 Jul 2017 09:11:13 PDT This dissertation explores how to attenuate the negative influence of team abusive supervision on followers’ creativity. First, it is not feasible to interfere with these negative impacts without the knowledge of the underlying processes. Essay 1 unravels how team abusive supervision gradually leads to followers’ creativity from a team process perspective by identifying team state mindfulness, team state negative affect, team task conflict, and team relationship conflict as critical mechanisms. Using a time-lagged multilevel field study of 92 teams, I found that abusive supervision aggravated relationship conflict via diminishing mindfulness levels in employees, and that abusive supervision exacerbated task conflict through elevating negative affect and decreasing team mindfulness levels. The augmented levels of relationship conflict in turn harmed individual creativity. However, followers’ creativity is not only influenced by team processes, but by leaders’ and followers’ individual differences as well. As such, essay 2 introduces leaders’ and followers’ individual differences (i.e., state positive affect) and examines how the state positive affect of both leaders and followers alters followers’ interpretations of leader mistreatment. Analyses of multilevel, multisource, and multiphase data show that leaders’ and followers’ state positive affect interactively determine the extent to which followers attribute abusive behaviors to their leaders’ performance promotion motives. Such attributions in turn benefit followers’ creativity. Implications and future directions are discussed.

]]> Yuanmei Qu http://scholarlyrepository.miami.edu/oa_dissertations/1913 http://scholarlyrepository.miami.edu/oa_dissertations/1913 Tue, 25 Jul 2017 09:11:09 PDT Malaria elimination remain a worldwide public health challenge. Residual malaria transmission has been consistently reported even in areas with adequate coverage of the two World Health Organization (WHO) recommended vector control interventions of long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS). Sugar-feeding and resting are two neglected yet crucial behaviors in malaria vector biology, and the corresponding environmental resources can play a huge role on the vector biology, and consequently on malaria transmission. However, there are few research on these two behaviors. The wide use of the indoor vector control tools has increased insecticide resistance in malaria vectors and the outdoor malaria transmission, which is a major cause for the failure of malaria elimination. Whether indoor vector control alone can achieve malaria elimination is unclear, and there are considerations of incorporating outdoor vector control interventions into the current integrated vector management. Novel tools such as attractive toxic sugar bait (ATSB), which can be used both indoors and outdoors, has been developed in response to the call from WHO. Its efficacy toward controlling malaria vectors has been proved by several field trials, however, more information on its epidemiological impact on malaria transmission and the optimum way to use it is needed. A spatial individual-based model was developed to 1) estimate the impact of environmental sugar sources and resting sites on the survival of malaria vectors and malaria parasite transmission; 2) evaluate the efficacy of ATSB stations and select the optimum number of spatial configuration of the stations for effective malaria control; and 3) evaluate the immediate and long-term effect of outdoor malaria vector control on malaria transmission. Results of study 1 show that small increase in the densities of sugar sources and resting sites resulted in significant increase of survival of malaria vectors and the human biting rates (HBRs). In addition, sugar sources dispersed over the whole area supported the survival of vectors better than being concentrated even near resting sites or houses. Results of study 2 show that ATSB application effectively reduced the density of malaria vectors, HBR and entomological inoculation rate (EIR) in both resource-rich and resource-poor environments. Configurations of dispersed ATSB stations were significantly more effective for vector control in resource-rich environments; but in resource-poor environments, all configurations worked similarly. Reduction of vector density and EIR increased with the increase in numbers of ATSB stations, but it reached a point at which further increase was ineffective. Results of study 3 show that the use of ATSBs as the outdoor intervention in combination with partial LLIN coverage had significantly better immediate and long-term impacts on vector control and EIR reduction than scaling up LLIN coverage to 100%. In conclusion, environmental sugar sources and resting sites play a crucial role in malaria transmission; ATSB stations dispersed over the whole area is recommended for the most effective vector control; and incorporating outdoor vector control into the current indoor interventions is recommended for malaria elimination.

]]> Lin Zhu http://scholarlyrepository.miami.edu/oa_dissertations/1912 http://scholarlyrepository.miami.edu/oa_dissertations/1912 Tue, 25 Jul 2017 09:11:03 PDT Notch signaling is involved in many physiological and pathological cell processes. The aberrant activation of this pathway leads to deregulation of cell cycle and abnormal proliferation, leading to tumorigenesis in breast, colon, pancreas, etc. Although the general mechanism of Notch signaling pathway has been well established, several aspects remain unknown about the assembly of Notch transcriptional activation complex, such as other co-factors that may be involved and the role of posttranslational modifications in regulating Notch signaling, etc. Our lab previously demonstrated that NACK is a critical co-activator of Notch signaling, that binds to the Notch1 ternary complex. Herein, my study demonstrated that p300 and CBP acetylate Maml1 on residues 188K and 189K to recruit NACK to the Notch1 transcription complex to drive the recruitment of RNA polymerase II at the initiation of transcription. Notably, NACK is recruited to the ternary complexes containing Maml1 and Maml3, but not Maml2, presenting the first evidence to support the hypothesis that Notch transcriptional co-factors have selectivity for specific Notch ternary complex. Moreover, simultaneous inhibition of p300 and Notch has a synergistic effect in esophageal adenocarcinoma. In summary, this study provides a deeper mechanistic understanding of the assembly of the Notch transcriptional complex and provides the rationale and proof of concept for a combinatorial therapeutic attack on Notch-dependent cancers.

]]> Ke Jin http://scholarlyrepository.miami.edu/oa_dissertations/1911 http://scholarlyrepository.miami.edu/oa_dissertations/1911 Mon, 24 Jul 2017 14:28:14 PDT Cardiovascular disease (CVD) is the leading cause of mortality globally, responsible for over 17.3 million deaths annually (~31% of all deaths). Over the next decade, this value is projected to increase, surpassing 23.6 million annual deaths by 2030. The vast majority of these deaths are caused by complications resulting from progressive thickening of the arterial wall, and the formation and growth of arterial plaques within the vessel lumen. Subsequent plaque rupture may potentially lead to blood clot (thrombus) formation. The combined effect of these processes is the reduction of vessel diameter (stenosis) and subsequent reduction in blood flow (ischemia), while complete occlusion of a vessel via thrombus formation may precipitate a heart attack or stroke. Utilizing discrete event simulation (DES) modeling, the various processes of lesion formation and growth were examined, and it was demonstrated that an increased number of raised lesions in an individual at an early age represents a reliable biomarker for increased risk of subsequent sudden cardiac death (SCD). As a result of this finding, the levels of known CVD risk factors for a group of ~3,000 young people (ages 15-34) were investigated. Using generalized linear regression modeling (GLM) the manner in which these factors contribute to lesion formation and growth was analyzed. From this analysis, a subpopulation (~13%) exhibiting an increased number of raised lesions was identified. However, no correlation between accelerated atherosclerosis and measured levels of risk factors was observed in this group, suggesting one or more “hidden” risk factors. Thus, a biosensor capable of identifying this “high-risk” group presenting a significantly increased number of lesions would be extremely beneficial. Atherosclerotic plaques possess a large number of apoptotic and necrotic cells. Based on this, a fusion protein was designed and expressed that was capable of targeting and binding to these apoptotic cells (Annexin V), as well as generating a detectable bioluminescent signal (RLuc8). This Annexin-Renilla Fusion Protein (ArFP) was subsequently characterized and applied both in vitro and in vivo, demonstrating that apoptosis detection is possible in a number of samples including vascular tissue. Plaque identification through apoptosis detection with this unique sensor provides another step towards the early identification and possible intervention of those at the highest risk of CVD-related death.

]]> Trajen Rex Head http://scholarlyrepository.miami.edu/oa_dissertations/1910 http://scholarlyrepository.miami.edu/oa_dissertations/1910 Fri, 21 Jul 2017 10:57:49 PDT The vertical wind shear measured between 200 and 850 hPa is commonly used to diagnose environmental interactions with a tropical cyclone (TC) and to forecast the storm's intensity and structural evolution. More often than not, stronger vertical shear within this deep layer prohibits the intensification of TCs and leads to predictable asymmetries in precipitation. But such bulk measures of vertical wind shear can occasionally mislead the forecaster. In the first part of this dissertation, we use a series of idealized numerical simulations to examine how a TC responds to changing the structure of unidirectional vertical wind shear while fixing the 200-850-hPa shear magnitude. These simulations demonstrate a significant intensity response, in which shear concentrated in shallow layers of the lower troposphere prevents vortex intensification. We attribute the arrested development of TCs in lower-level shear to the intrusion of mid-level environmental air over the surface vortex early in the simulations. Convection developing on the downshear side of the storm interacts with the intruding air so as to enhance the downward flux of low-entropy air into the boundary layer. We also construct a two-dimensional intensity response surface from a set of simulations that sparsely sample the joint shear height-depth parameter space. This surface reveals regions of the two-parameter space for which TC intensity is particularly sensitive. We interpret these parameter ranges as those which lead to reduced intensity predictability. Despite the robust response to changing the shape of a sheared wind profile in idealized simulations, we do not encounter such sensitivity within a large set of reanalyzed TCs in the Northern Hemisphere. Instead, there is remarkable consistency in the structure of reanalyzed wind profiles around TCs. This is evident in the distributions of two new parameters describing the height and depth of vertical wind shear, which highlight a clear preference for shallow layers of upper-level shear. Many of the wind profiles tested in the idealized simulations have shear height or depth values on the tails of these distributions, suggesting that the environmental wind profiles around real TCs do not exhibit enough structural variability to have the clear statistical relationship to intensity change that we expected. In the final part of this dissertation, we use the reanalyzed TC environments to initialize ensembles of idealized simulations. Using a new modeling technique that allows for time-varying environments, these simulations examine the predictability implications of exposing a TC to different structures and magnitudes of vertical wind shear during its life cycle. We find that TCs in more deeply distributed vertical wind shear environments have a more uncertain intensity evolution than TCs exposed to shallower layers of upper-level shear. This higher uncertainty arises from a more marginal boundary layer environment that the deeply distributed shear establishes, which enhances the TC sensitivity to the magnitude of deep-layer shear. Simulated radar reflectivity also appears to evolve in a more uncertain fashion in environments with deeply distributed vertical shear. However, structural predictability timescales, computed as the time it takes for errors in the amplitude or phase of azimuthal asymmetries of reflectivity to saturate, are similar for wind profiles with shallow upper-level shear and deeply distributed shear. Both ensembles demonstrate predictability timescales of two to three days for the lowest azimuthal wavenumbers of amplitude and phase. As the magnitude of vertical wind shear increases to universally destructive levels, structural and intensity errors begin to decrease. Shallow upper-level shear primes the TC for a more pronounced recovery in the predictability of the wavenumber-one precipitation structure in stronger shear. The recovered low-wavenumber predictability of TC precipitation structure and the collapse in intensity spread in strong shear suggests that vertical wind shear is most effective at reducing TC predictability when its magnitude is near the threshold between favorable and unfavorable values and when it is deeply distributed through the troposphere. By isolating the effect of the environmental flow, the simulations and analyses in this dissertation offer a unique understanding of how vertical wind shear affects TCs. In particular, the results have important implications for designing and implementing future environmental observing strategies that will be critical for improving forecasts of these destructive storms.

]]> Peter M. Finocchio http://scholarlyrepository.miami.edu/oa_dissertations/1909 http://scholarlyrepository.miami.edu/oa_dissertations/1909 Fri, 21 Jul 2017 10:57:45 PDT This comparative case study investigated integrated reading-writing instruction (IRWI) as an approach for writing instruction implemented after the adoption of Common Core State Standards (NGA & CCSSO, 2010a). Specifically, I explored the nature of IRWI through video-recorded observations of writing instruction, teachers’ perspectives towards this approach, and features of critical knowledge (Fitzgerald & Shanahan, 2000) utilized by five teachers in five fourth grade English Language Arts classrooms with emergent bilingual students. This study was guided by the following three research questions: 1) What is the nature of integrated reading-writing instruction in 4th grade ELA classrooms with emergent bilingual students? 2) What are teachers’ perceptions of integrated reading-writing instruction for emergent bilingual students? 3) What types of critical knowledge do teachers use when implementing integrated reading-writing instruction? This study analyzed videos collected within the frame of an Institute of Education Sciences-funded research-study, Writing for English Language Learners (WELLs): Exploring the Relationship between Writing Instruction and Student Outcomes (Gort, Howard, & Caswell, 2013). Additional data were also collected for this secondary study: teachers’ interviews and play-back sessions. The findings from this study indicated that the teachers’ instruction was built around three areas of the Florida Department of Education’s state assessment rubric: 1) Purpose, Focus, Organization; 2) Evidence and Elaboration; and 3) Conventions and Grammar. The investigation of classroom instruction showed how teachers paid particular attention to those aspects of writing that were explicitly stated in the assessment rubric. The five teachers saw a number of IRWI disadvantages, which included absence of creativity and genre variety, the approach not corresponding to a 4th grade developmental level, absence of an accurate picture of writing performance, and mismatch between assessment rubric and anchor papers. The teachers suggested a number of adaptations of IRWI, which, from their perspectives, would improve the current state of writing instruction, such as starting IRWI from kindergarten and throughout lower elementary grades, including more genre variety, writing in L1, among others. This study also contributed to the understanding of which elements of critical knowledge were present in the instruction of five teachers (Fitzgerald & Shanahan, 2000). The results of this study informed the field with regard to the five teachers’ practices and their perspectives of IRWI in 4th grade ELA writing classrooms with emergent bilinguals. This study built on, supported, and added to the existing research on writing instruction for emergent bilinguals in elementary school. Some important implications were suggested regarding what can be done to promote teachers’ implementation of IRWI to deepen their understanding of this approach.

]]> Irina Malova http://scholarlyrepository.miami.edu/oa_dissertations/1908 http://scholarlyrepository.miami.edu/oa_dissertations/1908 Thu, 20 Jul 2017 13:57:02 PDT Short-term load forecasting is crucial for the operations planning of an electrical grid. Forecasting the next 24 h of electrical load in a grid allows operators to plan and optimize their resources. The purpose of this study is to develop a more accurate short-term load forecasting method utilizing non-linear autoregressive artificial neural networks (ANN) with exogenous multi-variable input (NARX). The proposed implementation of the network is new: the neural network is trained in open-loop using actual load and weather data, and then, the network is placed in closed-loop to generate a forecast using the predicted load as the feedback input. Unlike the existing short-term load forecasting methods using ANNs, the proposed method uses its own output as the input in order to improve the accuracy, thus effectively implementing a feedback loop for the load, making it less dependent on external data. Using the proposed framework, mean absolute percent errors in the forecast in the order of 1\% have been achieved, which is a 30\% improvement on the average error using feedforward ANNs, ARMAX and state space methods, which can result in large savings by avoiding commissioning of unnecessary power plants. In addition, in order to improve the robustness of the forecast to variations in the number of neurons and other network parameters, the author proposes a method using an exponential decay of the error weights for training the neural network. The modification consists in giving higher error weight to more recent values and lower weight to older values of the training set. By doing this, mover recent values have a higher influence on the calculation of the synaptic weights and therefore the forecast produced by the NARX network is more accurate. This method, combined with the use of Bayesian regularization for training, results in improved forecast accuracy of up to 25\% and robustness to variation in parameter selection. The New England electrical load data are used to train and validate the forecast prediction.

]]> Jaime H. Buitrago http://scholarlyrepository.miami.edu/oa_dissertations/1907 http://scholarlyrepository.miami.edu/oa_dissertations/1907 Tue, 18 Jul 2017 15:38:07 PDT The knee meniscus is a complex structure that has been the focus of important and extensive research for many years. The most common areas of study include meniscal structure, meniscal tears, regeneration, and the biomechanical and anatomical forces endured by the meniscus. Although much research has been done regarding these tissues, more is needed in order to fully understand the role that menisci plays in the function and pathophysiology of the human knee. The fibrocartilaginous meniscus plays an essential role in distributing the majority of the load and maintaining not only congruency, but also lubrication in the knee joint. Degeneration of the knee meniscus is commonplace, yet its pathophysiology has not been fully explained. Because the meniscus is a nearly avascular tissue which lacks blood vessels for the delivery of nutrients, one area of study needing further research is the transport of fluids and solutes through meniscal tissues. In this dissertation, custom experimental methods are used to characterize the transport of solutes and fluids in meniscus fibrocartilage. For each study, we investigated the effects of mechanical strain, tissue anisotropy, and tissue region on the transport behavior in porcine meniscus tissue. Using a direct permeation experiment, hydraulic permeability was investigated to determine its strain- and/or direction-dependent behavior in porcine meniscus fibrocartilage. Our measured permeability values (1.53-1.87?10-15 m4/Ns) are similar to those in the literature for meniscus tissues. Results indicate that hydraulic permeability is anisotropic, being significantly greater in the circumferential direction than in the axial. Additionally, it was found that with increased compressive strain, there was a significant decrease in hydraulic permeability for all groups studied. Strain-dependent and anisotropic (i.e., direction-dependent) transport of glucose in porcine meniscus fibrocartilage was investigated using customized chambers to measure the diffusion and partition coefficients. Results indicate that both diffusivity and partitioning of glucose in porcine menisci significantly decrease with increasing compressive strain. Furthermore, diffusivity of glucose was found to be anisotropic, being significantly greater in the circumferential direction than the axial at all strain levels. Using the results from the partitioning and diffusion of glucose, we were able to calculate the effective diffusivity of porcine meniscus fibrocartilage. Finally, the strain-dependent and anisotropic electrical conductivity and relative ion diffusion was investigated in porcine meniscus fibrocartilage using a four-wire method. Results indicate that the conductivity and diffusion of ions in the meniscus significantly decreases with increasing compressive strain. Additionally, the conductivity and diffusion of ions was found to be significantly anisotropic, being greater in the circumferential directions than the axial direction at all strain levels. To our knowledge, this is the first study to quantitatively characterize the effects of strain, anisotropy, and region on transport properties in meniscus tissues. In particular, this is first study to measure glucose or ion diffusivity, glucose partitioning, or electrical conductivity in meniscus. The findings of this dissertation greatly enhance the knowledge of fluid and solute transport in the knee meniscus. Given that nutrient transport is critical for meniscus survival, this information can provide important insight into the functions and mechanisms of meniscus disease and even help identify effective treatment solutions for osteoarthritis.

]]> Kelsey L. Kleinhans http://scholarlyrepository.miami.edu/oa_dissertations/1906 http://scholarlyrepository.miami.edu/oa_dissertations/1906 Fri, 14 Jul 2017 10:57:51 PDT It is well established that Notch functions as a transcriptional activator through the formation of a ternary complex that comprises Notch, Maml and CSL. This ternary complex then serves to recruit additional transcriptional cofactors that link to higher order transcriptional complexes. The mechanistic details of these events remain unclear. This report reveals that the Notch ternary complex can direct the formation of a repressor complex to terminate gene expression of select target genes. Herein, it is demonstrated that p19Arf and Klf4 are transcriptionally repressed in a Notch-dependent manner. Furthermore, results indicate that Notch recruits Polycomb Repressor Complex 2 (PRC2) and Lysine Demethylase 1 (KDM1A/LSD1) to these promoters, which leads to changes in the epigenetic landscape and repression of transcription. The demethylase activity of LSD1 is a pre-requisite for Notch-mediated transcriptional repression. In addition, a stable Notch transcriptional repressor complex is identified containing LSD1, PRC2 and the Notch ternary complex. These findings demonstrate a novel function of Notch and provide further insight into the mechanisms of Notch-mediated tumorigenesis.

]]> Xiaoqing Han http://scholarlyrepository.miami.edu/oa_dissertations/1905 http://scholarlyrepository.miami.edu/oa_dissertations/1905 Thu, 13 Jul 2017 10:46:45 PDT Boosting is one of the most powerful machine learning method use for modeling a univariate response. However its application for the multivariate response is limited. We use gradient boosting approach (a generic form of boosting) for modeling multivariate response. Specifically we focus on the longitudinal data in which repeated measurements are observed for a subject over time. Our gradient boosting approach is use to boost multivariate tree to fit a novel flexible semi-nonparametric marginal model for longitudinal data. In this model, features are modeled non-parametrically using multivariate tree, while feature-time interactions are modeled semi-nonparametrically utilizing P-splines with estimated smoothing parameter. In order to avoid overfitting, we describe a relatively simple in sample cross-validation method which can be use to estimate the optimal boosting iteration and which has the surprising added benefit of stabilizing certain parameter estimates. Our new multivariate tree boosting method is shown to be highly flexible, robust to covariance misspecification and unbalanced designs, and resistant to overfitting in high dimensions. Feature selection is performed using variable importance to identify important features and feature-time interactions. We also explain some modification to the approach that improves the prediction performance. This includes using new gradient component as well as using random forest as the base learner. Additionally, we described a new multivariate boosting approach for the multivariate response when the data is generated from the cross-sectional study. In this approach, our aim is to detect covariates which are related to most of the response variables in the high-dimensional sparse setting. Throughout, the efficiency of our approach is demonstrated using simulated as well as real dataset.

]]> Amol Pande http://scholarlyrepository.miami.edu/oa_dissertations/1904 http://scholarlyrepository.miami.edu/oa_dissertations/1904 Wed, 12 Jul 2017 09:42:27 PDT Cholesterol (C27H46O) is the precursor of steroid hormones, Vitamin D and bile acid and is a key component of the plasma membrane. Cholesterol levels are tightly regulated through biosynthesis, cellular uptake or efflux pathways to remain within a certain range for normal cell survival and function. Disturbed cholesterol homeostasis can be an independent risk factor for human diseases including cardiovascular disease, neurodegenerative disorder and chronic kidney disease. Understanding the molecular mechanisms that govern cholesterol homeostasis has the potential to reveal important insights into disease pathogenesis and new therapeutics. Elevated plasma LDL cholesterol is a major risk factor for cardiovascular events and initiates the built up of atherosclerotic plaque in the arterial wall. In the arterial wall, the accumulated LDL cholesterol reacts with native oxygen species and becomes oxidized LDL that induces inflammatory responses and promote aortic Smooth Muscle Cells (SMCs) to undergo osteoblastic differentiation, proliferate, and hence contribute to the growth of atherosclerotic plaque. The first part of this study identifies microRNA-30e (miR-30e) as an OsteomiR in atherosclerosis. We present evidence that miR-30e regulates a panel of osteogenic genes in bone marrow derived Mesenchymal Stem Cells (MSCs), reduces osteogenic differentiation in aortic SMCs and the atherogenice ApoE-/- mice by directly repressing the expression of the osteogenic protein IGF2. MiR-30e also reduces calcification in SMCs. Our data reveals miR-30e-Igf2 as a novel pathway in osteogenesis-mediated atherogenic progression. In addition to causing cardiovascular disease, altered lipoprotein metabolism has also been reported in patients with chronic kidney disease. Although the exact molecular mechanism is not clear, retention of lipid in the renal epithelial cells has been thought to promote epithelial cells to undergo epithelial-to-mesenchymal transition that causes fibrogenic responses in the kidney. The second part of this study investigates the role of LDL-c in the renal tubules in an animal model of Alport Syndrome. Alport Syndrome, characterized by hereditary progressive renal dysfunction, deafness and visual anomalies is an inherited chronic kidney disease that is caused by mutations in type IV collagen α chains. Our study in the COL4A3-/- “Alport mouse” reveals that Osteopontin (OPN), is highly expressed in the renal tubules of the Alport mouse and that it plays a major causative role in Alport pathology by regulation of Dynamin 3 (DNM3) and LDL receptor (LDLR) in the renal tubules of Alport mice. Our results suggest a new pathway for Alport pathology where tubular OPN causes DNM3-mediated enhanced cholesterol influx. Our data suggest that OPN may be a druggable therapeutic target for Alport Sydrome.

]]> Wen Ding http://scholarlyrepository.miami.edu/oa_dissertations/1903 http://scholarlyrepository.miami.edu/oa_dissertations/1903 Wed, 12 Jul 2017 09:42:24 PDT Both depression and inflammation are independently associated with breast cancer health outcomes, and multiple studies have shown that depression and inflammatory markers may be linked among women with breast cancer. Studies of cognitive-behavioral based psychosocial interventions have found beneficial intervention effects on time to survival and recurrence in breast cancer patients. However, the mechanisms through which interventions affect clinical health outcomes are less understood. It has been suggested that psychosocial interventions may affect long-term breast cancer clinical disease endpoints via effects on immune and inflammatory processes, but more research is necessary to explore these relationships. The current study examined the relationships between post-surgical and pre-adjuvant levels of depressive symptoms and pro-inflammatory cytokines with long-term clinical health outcomes, both individually and in combination, among a cohort of non-metastatic breast cancer patients. It also sought to replicate recent findings that a cognitive behavioral stress management (CBSM) psychosocial intervention predicts favorable breast cancer clinical disease endpoints, and examined possible mediators of these effects. The present sample included 90 women with non-metastatic breast cancer and available blood data for analyses of serum pro-inflammatory cytokines from a larger clinical trial. Women were initially recruited and assessed at 2-10 weeks post-surgery (T1). At that time, information was collected related to demographic characteristics, medical history, treatment plans, and psychosocial functioning. Women were randomized to either a 10-week group-based CBSM intervention or a 1-day psychoeducational group seminar control. Participants were re-assessed at 6 months (T2), 12 months (T3), and 5-years (T5) post-T1. Blood samples for cytokine analyses were collected at T1 and T3. At 8-15 year follow-up (11-year median; T6), a tumor registry linkage was performed and medical chart reviews were conducted to determine mortality status (including cause and date of death) and disease free status (i.e., recurrence status) of study participants. Cox Proportional Hazards Models were conducted to assess the direct effects of baseline depressive symptoms and serum concentrations of pro-inflammatory cytokines (i.e., IL-1β, IL-6, and TNF-α) on time to clinical disease endpoints (i.e., all-cause mortality, breast cancer-specific mortality, and breast cancer recurrence) assessed at 11-year median follow-up (Aim 1a). Bootstrapped linear regression analyses were used to test indirect relationships between baseline depressive symptoms and pro-inflammatory cytokines with time to clinical disease endpoints (Aim 1b). Cox Proportional Hazards Models were conducted to assess for group differences (i.e., CBSM vs. control) in time to clinical disease endpoints (Aim 2a). Finally, bootstrapped linear regression analyses were used to test indirect effects of CBSM on time to clinical disease endpoints through 12-month changes in depressive symptoms and pro-inflammatory cytokines (Aim 2b). Unadjusted and adjusted models were conducted, which accounted for age, stage of disease, surgical procedure, hormone therapy, and smoking status. All analyses were run in the 90 women for whom blood data were available, and in a subset of 73 women who initially had invasive disease (i.e., not stage 0). At median 11-year follow-up, 8 women (8.9%) were deceased and 6 of those deaths (75.0%) were related to breast cancer. A total of 17 women (18.9%) had experienced a breast cancer recurrence. In Aim 1a, results of Cox Proportional Hazards analyses revealed non-significant relationships between baseline variables and time to clinical disease endpoints in both the full and invasive subsamples in unadjusted and adjusted models (all ps > 0.10). In Aim 1b, unadjusted and adjusted linear regression analyses revealed significant associations between greater baseline depressive symptoms and concurrent greater serum concentrations of IL-1β (β = 0.29, p = 0.007) and TNF- α (β = 0.30, p = 0.004). A marginal association emerged between greater baseline depressive symptoms and concurrent greater IL-6 (β = 0.179 p = 0.077). These findings were retained in the invasive subsample. However, as in Aim 1a, baseline levels of depressive symptoms and pro-inflammatory cytokines did not predict time to clinical disease endpoints (all ps > 0.10), and thus mediation was not supported. In Aim 2a, results of Cox Proportional Hazards analyses revealed non-significant group differences in time to clinical disease endpoints in both the full and invasive subsamples in unadjusted and adjusted models (all ps > 0.10). In Aim 2b, results of linear regression analyses revealed non-significant group differences in 12-month changes in depressive symptoms and pro-inflammatory cytokines (all ps > 0.10), and mediation was therefore not supported. The observed associations between baseline depressive symptoms and pro-inflammatory cytokines have implications for the treatment of women with breast cancer who report comorbid elevated depressive symptoms. However, the long-term implications of these findings, including the role of psychosocial interventions, are inconclusive. More research is needed, including large well-controlled trials, to further investigate the associations between these variables to elucidate the mechanisms through which depressive symptoms, inflammation, and psychosocial interventions interact and ultimately affect long-term clinical health outcomes of breast cancer patients.

]]> Laura Bouchard http://scholarlyrepository.miami.edu/oa_dissertations/1902 http://scholarlyrepository.miami.edu/oa_dissertations/1902 Wed, 12 Jul 2017 09:42:20 PDT Nowadays people spend 90% of the time in indoor. To provide a comfortable and healthy environment for occupants, buildings consume 40% of the total energy in the world. Due to the inappropriate design of the indoor environment, the problems related to bad indoor air quality caused over $20 billion loss in the US. Then it raises a question on how to improve the indoor environment and decrease the energy consumption in the buildings. One of the strategies available is to utilize the stratified airflow distributions such as mixed mode ventilation. Previously, the coupled simulation between building energy simulation program and computational fluid dynamics (CFD) models was used to study energy and comfort performance of those systems while putting the control dynamics aside. This research develops the coupled simulation model to study the dynamic systems of non-uniform airflows, HVAC, control, and building envelopes. Fast fluid dynamics (FFD) is chosen to simulate non-uniform airflows since FFD is computationally fast than FFD in simulating the non-uniform airflows. Modelica language, which is equation-based and object-oriented, is used to model HVAC, control, and building envelopes. Then, the coupled simulation model is further ameliorated by adding the multizone models to expand the application scope of the model from a single zone to a building with multi zones. To further improve the model for design optimization study, this research improves the computation speed of FFD by parallelizing it using open computing language (OpenCL). We systematically evaluated on the feasibility of using OpenCL to accelerate the airflow simulation using FFD as an example. Though FFD programmed in OpenCL running on different graphics processing units (GPU) may generate different results due to different interpretation of IEEE-754 standards, the difference is minor to some extents that are negligible in airflow simulation. Running FFD in parallel on a, up to 1139 times speedup is achieved, which is promising to dramatically reduce the time cost for design optimization of the dynamic systems. Regarding the operation optimization, it would be preferable to increase the computation speed of non-uniform airflow simulations by using reduced order models (ROM). We proposed to use in situ adaptive tabulation (ISAT), which differentiates from other conventional ROMS in that it can call a full-scale FFD simulation when the prediction is deemed not accurate. This is critical in the optimization. ISAT executes a FFD simulation if interpolation is deemed inaccurate. In this study, ISAT is trained by using FFD running in parallel on a GPU and once well trained ISAT can answer query points both inside and close to training domain using retrieve actions within a time less than 0.001s for each query. This shows that ISAT can be used to further improve the coupled simulation model to realize operation optimization, such as model predictive control using a non-gradient based optimization.

]]> Wei Tian http://scholarlyrepository.miami.edu/oa_dissertations/1901 http://scholarlyrepository.miami.edu/oa_dissertations/1901 Tue, 11 Jul 2017 07:38:42 PDT Development of resistance and relapse remains a major challenge for acute lymphoblastic leukemia (ALL) treatment. Results from our lab and others showed that ALL cells are specifically vulnerable towards disturbance of endoplasmic reticulum (ER) protein homeostasis, which will cause proteotoxic ER stress/unfolded protein response (UPR). The typical ER stress inducer bortezomib (VELCADE) has been approved as a front-line treatment for multiple myeloma. Clinical trials of bortezomib in combination with other therapeutic agents demonstrated its potent anti-leukemic efficacy in relapsed/refractory pediatric ALL patients. Bortezomib functions by inhibiting the proteasome in the ubiquitin-proteasome system (UPS) to impede protein turnover in cancer cells, rendering accumulation of abnormal and/or misfolded proteins to trigger proteotoxic ER stress. However, the ubiquitous existence of proteasome in normal tissue cells compromised the selectivity of this therapeutic strategy as severe side effects have been reported in clinical trials with bortezomib treatment. Nonetheless, the great success of bortezomib in the clinic inspired the development of novel pharmacological strategies focused on targeting upstream components of the UPS. Currently, more than 1000 E3 ligases have been identified and the largest family among them is the cullin-RING ligases (CRLs) which are responsible for degradation of approximately 20% of cellular proteins degraded via UPS. Most of these proteins are critically involved in cell cycle progression, signaling pathway transduction and apoptosis induction. Activation of CRLs requires conjugation of an ubiquitin-like protein NEDD8 (neural precursor cell-expressed developmentally downregulated 8) to near the C-terminal of the scaffold cullin proteins in the CRLs. Consequently, NEDD8 conjugation serves as an upstream regulatory mechanism that can switch "on" and "off" CRL activity by NEDDylation and deNEDDylatin of cullins, respectively. Conjugation of NEDD8 to the cullins happens in three enzymatic steps involving NEDD8 activating enzyme (NAE; E1), UBC12 and UBE2F (E2s) and E3s that will carry out the final step in conjugating NEDD8 to cullins. A novel therapeutic agent called pevonedistat (pevo, MLN4924) was recently developed to specifically inhibit the attachment of NEDD8 to NAE, leading to blockage of NEDD8 conjugation to the cullins and the concomitant accumulation of CRL substrates. Based on the promising in vitro data proving the efficacy of this novel targeting strategy, multiple clinical trials are undergoing to evaluate the safety and efficacy of pevo in melanoma, acute myeloid leukemia and solid tumors. To evaluate the potential therapeutic gain of NEDDylation inhibition for ALL management, our study focused on analyzing the in vitro and in vivo efficacy of pevo in ALL, investigated potential mechanisms of action and studied the potential resistance ALL cells may develop with pevo treatment. Potent in vitro anti-leukemic activities of pevo were observed with IC50 (growth inhibition) and EC50 (cell death) around 200 nM and 400 nM, respectively, for the ALL cell lines studied (T-ALL: Jurkat, CCRF-CEM; Pre-B ALL: NALM6, REH, SUPB15). Mechanistic studies on the cytotoxicity of pevo identified concomitant activation of the mTOR/p70S6K cascade and eIF2a de-phosphorylation which will promote nascent protein synthesis to trigger proteotoxic ER stress/UPR as manifested by accumulation of GRP78 (Ig heavy chain-binding protein (BiP)) and CHOP (CCAAT/enhancer binding protein homologous protein) in pevo-treated ALL cells. Inhibition of nascent protein synthesis using the mTOR inhibitor rapamycin or the general protein synthesis inhibitor cycloheximide rescued pevo-induced proteotoxicity and ALL cell death, indicating pevo-induced ER stress/UPR as one major cytotoxic mechanism of NEDDylation inhibition in ALL cells. We identified CReP as a target of CRL that will accumulate with pevo-mediated NEDDylation inhibition and promotes eIF2a de-phosphorylation. Pevo induced eIF2a de-phosphorylation will compromise the ability of ALL cells to stop protein synthesis and thus, dysregulate the activation of cytoprotective integrated stress response (ISR), leading to sensitization of ALL cells towards chemotherapeutic agents like dexamethasone and others. We also found consistent activation of the MEK/ERK pathway in pevo-treated ALL cells. Blockage of the MEK/ERK pathway significantly potentiated the ALL-killing effect of NEDDylation inhibition in vitro and in vivo. We further identified Ca2+ influx via the Ca2+ release-activated Ca2+ (CRAC) channels activated protein kinase C fl (PKC-fl) which will promote MEK/ERK signaling cascade activation. The pro-survival role of MEK/ERK signaling in ALL cells were attributed to ERK1/2 mediated phosphorylation/neutralization of the pro-apoptotic Bcl-2 family protein BIM. Studies on the effects of NEDDylation blockage on the two essential components of the CRACs, Orai1 and stromal interaction molecule 1 (STIM1) showed that pevo treatment significantly decreased the STIM1 level while keeping Orai1 levels stable. These changes will optimize the Orai1:STIM1 ratio to activate CRAC, leading to Ca2+ influx in pevo-treated ALL cells. Further studies identified a unique upstream open reading ORF (uORF) in the 5'-untranslated region (5'-UTR) of STIM1 mRNA which can regulate STIM1 mRNA translation with phosphorylation and de-phosphorylation of eIF2a. Based on these data, the efficacy, cytotoxic mechanisms and drug-drug interactions with other chemotherapeutic agents of pevo-mediated NEDDylation inhibition in our pre-clinical ALL cell line models and NSG leukemia mouse model were identified, supporting NEDDylation blockage as a novel therapeutic strategy for ALL treatment.

]]> Shuhua Zheng http://scholarlyrepository.miami.edu/oa_dissertations/1900 http://scholarlyrepository.miami.edu/oa_dissertations/1900 Thu, 06 Jul 2017 09:07:58 PDT Young Spanish-speaking Latinos in the U.S., most of whom are from low-income backgrounds, perform below their English-speaking peers at kindergarten entry. This achievement gap is concerning considering the rising number of Latino children in the U.S. living in poverty. Despite this risk, a large body of research highlights the positive effects of learning two languages. Latino DLLs attending Head Start, compared to their monolingual peers, were recently found to have higher executive functioning (EF), a set of domain-general cognitive skills that robustly predict academic achievement. This emerging evidence is encouraging, however, there is still a lack of research on how this bilingual-EF advantage contributes to young DLLs’ school readiness in the context of early education classrooms. To better understand these factors, this study examined bilingual language, EF, and science achievement across the year in a sample of 424 Latino preschool DLLs across 38 Head Start classrooms. Children were assessed in the fall and spring on all measures, and observations of Spanish and English support in the classroom were conducted in the winter. Results from a cross-lag model demonstrated a significant bidirectional relationship between bilingual ability and EF across the year, and also indicated positive effects of both constructs on children’s science at the end of the year. Spanish and English support in the classroom did not influence the cross-lag paths between bilingual ability and EF across the year, however, English support appeared to moderate children’s EF from fall to spring, and Spanish support predicted both bilingual ability and EF at the end of the year. Results from this study help inform the mechanisms behind the bilingual-EF relationship and demonstrate positive effects on achievement. Additionally, findings highlight the importance of supporting English and Spanish for DLLs in the early childhood classroom.

]]> Lisa J. White