Methods from dynamical systems theory are applied to examine flow structures resulting from the interaction of actuated elastic structures with a surrounding fluid. In particular, finite-time Lyapunov exponent (FTLE) fields are employed in order to gain geometric insight to flow structures of biological relevance. Ridges of the FTLE fields correspond to Lagrangian coherent structures (LCS), which are fluid-fluid boundaries separating fluid domains into regions of qualitatively different flow. We begin by studying the flow in the periciliary region of the upper respiratory tract of the lungs generated by a motile, internally-actuated cilium. An integrative model that couples the force-generating mechanics of a cilium with the external fluid mechanics is examined. The computed LCS uncovers a fluid boundary that separates the fluid that gets advected downstream from the fluid that recirculates near the cilium, which arises from the asymmetric beat form and gives rise to complex mixing In many biological settings, fluid that is pumped peristaltically has non-Newtonian responses. We compare well-known geometric flow structures due to peristaltic pumping of a Newtonian fluid to those in a viscoelastic fluid, both in the case of a periodic channel and in a cavity with closed ends. The presence of sidewalls introduces a return flow that must result due to volume conservation. Cellular flow patterns are observed for Newtonian and Oldroyd-B fluids Finally, we analyze a time-periodic model of pulmonary airway reopening, modelled as a pulsatile finger of air in a fluid-occluded tube. The computed LCS uncovers flow structures which separate fluid that gets pushed downstream from fluid that flows into the upstream thin-film region. In certain flow regimes, a third region which moves along the bubble interface both upstream and downstream is observed / acase@tulane.edu

The purpose of this investigation is to determine and to detect signs and patterns of power system dynamic behaviors that lead to voltage instability long before approaching the point of system voltage collapse In large scale power systems, voltage instability or voltage collapse is known to be a 'slowly' occurring phenomenon that result when power systems operate very close to their transmission line operating limits, are forced to transmit power over along distance, or the system has insufficient static and dynamic reactive power. Because causes of voltage instability are variations in load demand, P-V or V-Q curves of certain buses or the entire system constitute the basis of voltage instability scenarios. Although voltage collapse is thought of involving dynamic variation of 'voltage magnitudes', to certain degrees, they also include changes in 'voltage angles' from their transient stable equilibrium points. Most voltage collapses in large-scale power systems include a combination of 'voltage Angle' and 'Voltage Magnitude' changes, and under heavy loading of the systems, it is hard to decouple the two system behaviors from each other In this investigation, we use system models that are suitable for real-time simulation and observe dynamic behavior of power systems in response to changes in system loading, application of transmission system contingencies and disturbances, and change in the mixture and availability of static and dynamic reactive power availability in the system. Using the proposed system changes as system stimulus, we observe and record dynamic behavior of the system in both time and frequency domains. Methods of artificial intelligence, pattern recognition, template analysis, frequency spectrum analysis, eigen value analysis, time response, and traditional P-V and V-Q methodologies are used in our modeling, simulation, and analysis of power systems To simulate small and large-scale systems, we use EMTP/SGI real time simulation capabilities. The effort is conducted using complementary real time simulation capabilities of Tulane University and Entergy Services, Inc. In addition to 10-Bus test system, we will simulate a 39-Bus test systems and a real-size transmission system that is created based from Entergy's transmission system / acase@tulane.edu

During the development of neuronal networks, ionotropic glutamate receptor signaling regulates both activity-dependent morphological plasticity and homeostatic synaptic plasticity. Dependent upon the strength of the afferent glutamatergic input, the interplay between these two processes likely determines the activity features of the mature network. However, it has not been shown whether the relative strength of the glutamate signal leads to proportional and relative plasticity-mediated adaptions in developing networks. Here, I show that by blocking activity prior to synaptogenesis, with two proportional levels of a combination of ionotropic glutamate receptor antagonists, and examining after putative synaptic maturation, both groups exhibited significant recovery of some activity-related parameters. These parameters included an increase in the incidence and frequency of Ca2+ transients, an increased amplitude of Ca2+ responses to bath applied glutamate, an increase in intrinsic excitability and the ability to induce non-NMDA receptor-mediated synchronized bursting upon disinhibition. Concomitantly, when removed from their respective concentrations of antagonists, I found proportional regulation of Ca2+ responses to bath applied glutamate and a tending towards a proportional increase in mEPSC frequency. In addition, I found proportional regulation of the AMPA subunit, GluR1 expression, the NMDA receptor subunit, NR1 expression and the expression of the beta-subunit of CaMKII. This work suggests that the activity properties of mature cultured hippocampal networks are function of the relative ionotropic glutamate signal through the period of synaptogenesis / acase@tulane.edu

The management of a variety of distributed resources for optimum energy efficiency could become an enabling technology to supplement the current energy system. Wind generators, micro-generators, photovoltaic generators, and fuel cells can be appropriately integrated to provide on-site power to consumers. For an increased system reliability and availability of a distributed supply system during outages of weather dependent sources, several energy storage devices may be employed in the energy conversion system. Consequently, a flexible control system needs to be developed to harness an optimal amount of power from a variety of sources when operating under both steady-state and transient conditions. In this dissertation, a novel high frequency multistage power conversion system is investigated for providing efficient energy transfer from a variety of sources to static, dynamic, linear, and nonlinear loads. The proposed multistage conversion system employs several power converters, including a six-pulse AC-DC rectifier, which contributes to the nonlinearity of the system, resulting in an increased harmonic distortion in the source current. Therefore, an attractive alternative to the six-pulse AC-DC rectifier is desired for the proposed energy conversion system Multipulse converters offer a substantially reduced value of distortion in the source current, making them potential alternatives for the standard front-end AC-DC rectifier. Multipulse converters, which consist of several series or parallel-connected rectifiers, produce a multistep source current as opposed to the three-step current offered by a six-pulse rectifier. Further reduction in the distortion can be achieved by adding more rectifiers, resulting in a higher pulse operation. However, an increase in the pulse number compromises the converter simplicity and cost due to complex transformer winding arrangements For improved performance and reduced component count, several hybrid multipulse converter topologies have been proposed. One investigated configuration employs an active current shaper, termed as the load compensator (LSTATCOM), for both the standard 12-pulse diode and thyristor converters. It is demonstrated that the LSTATCOM-based converter provides sinusoidal shaped source currents over wide variations in the output load and also compensates a limited amount of sag and swell in the source voltage / acase@tulane.edu

There are many diseases in the eye including glaucoma, macular degeneration, and diabetic retinopathy in which early detection is crucial in order for treatments to be effective. Current detection methods are invasive and costly which lead to many of these diseases going undiagnosed. If not caught early, these diseases can seriously impair vision and even lead to complete blindness. The overreaching goal of our project is to develop a system that can detect these diseases early as well as help further understand the mechanisms that cause these diseases. To accomplish this goal, a snapshot multispectral imaging system was developed that is capable of capturing seven wavelengths simultaneously for the purpose of calculating oxygen saturation (OS). The main specific aim of this project is to determine if our system can detect changes in OS by capturing images of hemoglobin in vitro. It was found that there is a linear relationship with strong correlation (R2=0.91) between the OS our system detected and verified OS percentages from a blood oximeter. A linear equation was found from the data set that has a slope of 0.25 and y-intercept of -0.035. In this study, the stability of the system was also looked at by quantifying the variance between images of the same patient. A maximum of 4% error was found between images which is similar to other systems in the field. We also wanted to determine whether our in vitro correlation could be used to calculate OS percentages in vivo. The results were inconclusive because of a 23% standard deviation in the data but the hypothesis is that the in vitro correlation cannot be used directly in vivo because the backgrounds are different. The results of the main specific aim of this study proved that the concept of our design works in vitro which gives us confidence to continue to work towards a system that is clinically useful in vivo / acase@tulane.edu

Temporal lobe epilepsy (TLE) is a common neurological disorder that is characterized by changes in the neuronal circuitry that make the brain susceptible to seizure generation. Cannabinoid type 1 (CB1R) and vanilloid (TRPV1) receptors are widely distributed in the brain including the temporal lobe. Neurons are capable of making endogenous cannabinoids called endocannabinoids that can act on CB1R and TRPV1. Recent studies implicate a role for cannabinoids in the treatment of epilepsy. However, direct cellular effects of CB1R and TRPV1 activation in an epileptic brain that has undergone synaptic reorganization have been understudied. These experiments were aimed at studying the effects of CB1R and TRPV1 activation in the dentate gyms granule cells of a murine model of TLE that has undergone synaptic reorganization In mice with mossy fiber sprouting and spontaneous seizures following pilocarpine-induced status epilepticus, cannabinoid agonists suppressed excitatory synaptic activity. This effect was blocked by the selective CB1R antagonist, AM-251. In these mice, the TRPV1 agonist capsaicin increased excitatory synaptic activity and this increase was blocked by the selective TRPV1 antagonist capsazepine. Anandamide, an endocannabinoid, increased excitatory synaptic activity when CB1R were blocked with AM251 but suppressed the activity when TRPV1 were blocked with capsazepine. CB1R and TRPV1 protein levels were also increased in these animals. The data presented here support the hypothesis that there is an effective redistribution of CB1R and TRPV1 in the dentate gyms of epileptic mice and that activation of these receptors decrease and increase synaptic activity respectively. These studies elucidate the interaction between cannabinoid and vanilloid systems that is essential for the development of new therapeutic strategies for treating TLE based on the cannabinoid system / acase@tulane.edu

Sweet almond beta-glucosidase was used to catalyze the hydrolysis of paranitrophenol-beta-D-glucopyranoside (pNPG) and methyl-beta-D-glucopyranoside (methyl-glc). The second order rate constants (kcat/Km) were determined over the range of pLs (pL = pH or pD) at which the enzyme is active, resulting in bell shaped plots of rate constants versus pL. k cat/Km represents the mechanistic steps up through and including the first irreversible step (leaving group dissociation). The second order limiting rate constants (kcat/Km)lim show a small solvent kinetic isotope effect (SKIE H2O(V/K) lim/D2O(V/K)lim), 1.15 (+/- 0.03) with pNPG and 1.05 (+/- 0.08) with methyl-glc. The observed acid dissociation constants (pKas) of the enzymic carboxylic acids for methyl-glc are ∼1 pH unit more acid than those for pNPG. These results support that enzyme catalyzed hydrolysis of methyl-glc hydrolysis is dependent on the pKa of an additional anionic residue The inhibitory effects of glucose on the enzyme catalyzed reaction of pNPG were monitored in H2O and D2O. Glucose showed negligible change in binding affinity with pH or substitution of D2O for H 2O, indicating no change in the ability of the glycon moiety of the substrate (and possibly the substrate itself) to bind to the enzyme. The shifts seen in kcat/Km are representative of a shift in k cat A new method of determining the rate of hydrolysis of glucono-1,5-lactone is introduced. The method is applicable at pHs 6 to over 7, where the rate of hydrolysis depends on solvent conditions. The assay is a low cost, efficient and easy method to obtain the rate of hydrolysis / acase@tulane.edu

Part I, Catalytic Mechanism of beta-Glucosidase. Sweet almond beta-glucosidase is a highly efficient family 1 exo-glycohydrolase. The activity of this enzyme depends on a deprotonated carboxylate and a protonated carboxylic acid for the optimal activity. In this study we report on the hydrolysis of O/S-glycoside catalyzed by beta-glucosidase. While the Km values for the S- and O-glycosides are similar, the kcat values are about 1,000-times lower for the S-glycosides. Remarkably, the pH-profile for kcat/Km for hydrolysis of p-nitrophenyl thioglucoside (pNPSG) shows the identical dependence on a deprotonated carboxylate and a protonated groups does the pH-profile for hydrolysis of the corresponding O-glycoside. There is no solvent kinetic isotope effect on the enzyme-catalyzed hydrolysis of pNPSG, so is pNPG. These results suggest the reason that the glutamic acid residue must be protonated is not that it is acting as a general acid catalyst, but rather, that the unionized glutamic acid avoids electrostatic repulsion with the incipient thiolate. Linear free energy relationships between the binding and catalytic parameters and the aryl substituents on the phenyl glucosides reveal a striking similarity between the S-and O-glycosides. For both substrates, catalysis (kcat/K m) shows a strong dependence on the pKa of leaving group (betalg = -1) suggesting that in the transition state there is nearly full development of a negative charge on the (thio)phenolic oxygen or sulfur. Substrate binding shows a dependence on both hydrophobicity and the electronic properties of aryls substituents. The Bronsted coefficient for substrate binding is beta ≈ -0.41, suggesting a weakening of glucosidic bond when the E·S complex is formed Part II, Protein Chemistry of beta-Glucosidase. As a glycoprotein, beta-glucosidase from sweet almond still has no crystal structure with around 5% carbohydrate. Endoglycosidase H but not peptide N-glycosidase F, could cleave carbohydrate from beta-glucosidase. This results suggest that the oligosaccharides of beta-glucosidase are attached to an N-terminal asparigine or C-terminal asparigine. With substrates pNP-Glc and pNP-Fuc the catalytic rate by enzyme-treated beta-glucosidase didn't change comparing to the native enzyme. While with substrate pNP-Gal the catalytic rate increased, with substrate cellobiose the catalytic rate decreased. Therefore it is possible that carbohydrate does play a role in the process of enzymatic catalysis beta-Glucosidase is a stable enzyme that only loses 50% activity after incubation with 8 M urea for 1 hour. beta-glucosidase can be unfolded and also inhibited by urea. Modification of beta-glucosidase by DTNB and NTCB did not change the enzymatic hydrolysis activity. Therefore, this interaction would not change the conformation around catalytic center. At least the thiols groups were not required for this enzymatic catalysis / acase@tulane.edu

Clustering and Classification are two of the most common data mining tasks, used frequently for data categorization and analysis in both industry and academia. Clustering is the process of organizing unlabeled objects into groups of which members are similar in some way. Clustering is a kind of unsupervised learning algorithm. It does not use category labels when grouping objects. In Semi-Supervised clustering, some prior knowledge is available either in the form of labeled data or pair-wise constraints on some of the objects. Classification is a kind of supervised learning algorithm. It is a procedure to assign class labels. A classifier is constructed from the labeled training data using certain classification algorithm, it then will be used to predict the class label of the test data In this dissertation, the results of a comprehensive comparative study of three kinds of clustering algorithms including Co-Clustering, Consensus-based Clustering and Semi-supervised Clustering is presented. Through experiments using artificial datasets with different data substructures and UCI data sets, the performance of the three kinds of clustering algorithms was compared and analyzed. A method was proposed to combine a Co-Clustering algorithm and a Semi-supervised Clustering algorithm. A comprehensive comparative study was conducted on three kinds of classification algorithms including Logistic Regression Classifier, Support Vector Machine and Decision Tree. Experiments were carried out using different artificial datasets and UCI data sets to analyze and compare their classification performance. A method using controlled False Discovery Rate was proposed in Logistic Regression Classifier to select important features. A detailed proof was developed to show that controlling False Discovery Rate can be achieved by controlling the related p-value. Experiments were also conducted to compare the classification performance using the proposed feature selection algorithm Keywords. Classification, Clustering, Semi-supervised Clustering, Feature Selection / acase@tulane.edu

In this thesis, we provide a starting point for the development of computational models of both uterine and oviductal fluid mechanics. The method of regularized Stokeslets provides a framework that is easily implemented and may be adapted to include the full fluid-structure interactions that are observed physiologically We compare our computational results to analytic results in settings where the geometry of the channel (in 213) or tube (in 3D) undergoes small amplitude, symmetric contractions. These comparisons allow us to validate our numerics, and give us insight into the choice of our numerical parameters We demonstrate that our results for tapered channels and non-symmetric channels in two dimensions match the lubrication theory of Elad et al. [6]. We extend these simulations to three dimensions. In addition, we include an inserted catheter in our two-dimensional model, as in the finite element calculations presented by [25]. Finally, we present preliminary results of a full, three-dimensional model of a peristaltic tube with an inserted catheter / acase@tulane.edu