Group A Streptococcus (GAS) is responsible for mild and common infections like tonsillitis and pharyngitis, and more serious invasive disorders like necrotizing fasciitis and glomerulonephritis. The ability to invade tissues is closely linked to the virulence factors expressed by the bacterium. Hyaluronic acid capsule expression is variable among all the strains in S. pyogenes and confers the capacity to evade the immune response. In a previous study, it was found that capsule production in CovR mutants was temperature-regulated, showing a basal level of capsule production at 37℃ but increased production was observed at 25℃. Moreover, it was found that when CvfA, an endoribonuclease, is mutated, this thermoregulation is abolished. Since an antisense RNA was found spanning the entire capsule operon, another ribonuclease, RNaseIII which is involved in dsRNA digestion, was considered as a possible candidate to regulate the antisense and capsule transcripts. In this study, the objective is to find if and how the ribonucleases regulate the capsule transcript and antisense RNA. Data suggests that RNaseIII does not have a role in the regulation. On the other hand, CvfA showed a key role in regulating transcript levels. Furthermore, temperature appears to have some effect on its activity.

Posterior capsule opacification (PCO) is a major component of secondary cataract, a complication of current cataract surgery practice. This iatrogenic condition occurs in virtually all pediatric cases and to a lesser extent in adults. PCO correlates with the development in the latter half of the 20th Century of extracapsular cataract extraction (ECCE). In these surgeries, the lens capsule is left intact. During ECCE surgery a circular capsulotomy opening is created in the anterior lens capsule, and the cataractous, proteinaceous lens is removed, often via ultrasonic lens liquefaction i.e. phacoemulsification. The posterior, equatorial and remaining anterior portions of the sac-like capsule are left intact, permitting the insertion of an artificial lens into the emptied capsule. However, cells from the monolayer of epithelium on the inner surface of the capsule often begin to proliferate and migrate onto the normally cell-free inner surface of the posterior capsule, and may obscure the central axis of vision. Subsequently, a second surgery is necessary to create a small capsulotomy in the centre of the posterior capsule, usually employing an Nd:YAG laser. However, up to 5% of patients who have capsulotomies may then develop further serious, vision-threatening complications such as macular edema and retinal detachments. This thesis reports the photodynamic therapy (PDT) conditions required to prevent lens epithelial (LE) cell de novo proliferation and migration onto posterior lens capsules in a euthanized rabbit surgical model in order to predict parameters required to prevent PCO in humans. Experiments with primary in vitro cultures of human LE cells have shown rapid delivery of the photosensitizer benzoporphyrin derivative monoacid ring A (BPD-MA) and efficient killing with low light doses of 690 nm red light. Additional studies have shown the efficacy of various viscous agents in protecting the comeal endothelium. During model phacoemulsification ECCE surgeries, the use of hyaluronate viscoelastic carriers addressed the need for containment necessary for localized delivery of photosensitizer in the emptied capsule. Long-term monitoring of PDT-treated rabbit lens capsules in vitro has demonstrated a phototoxic effect including complete cell kill in this surgical model employing the prophylactic use of PDT. / Medicine, Faculty of / Pathology and Laboratory Medicine, Department of / Graduate

Campylobacter jejuni is a leading cause of food poisoning and according to the World Health Organisation accounts for majority of the 4.5 billion cases of global food poisoning each year. Genome sequencing by Parkhill et al. (2000) identified a gene, cj1411c, which is thought to encode a lone cytochrome P450, CYP172A1. In this thesis the role of CYP172A1 was studied using in vivo and in vitro techniques. The genomic location of cj1411c is adjacent to the capsular biosynthetic genes. The capsular and P450 genes are conserved in some species of Campylobacter and Helicobacter, as well as in Comamonas testosteroni. Importantly, this work has demonstrated that the P450 gene is expressed in two well characterised laboratory C. jejuni strains, 11168H and 81-176. Protein production was disrupted using insertional knockout mutagenesis, which allowed for investigations into the role of the enzyme in the host. Alterations to the observed autoagglutination rate and growth characteristics indicated that CYP172A1 has a role in modifying the bacterial surface. The insertional knockout mutant also resulted in cells which were more susceptible to detergent-like compounds (e.g. polymyxin B and sodium deoxycholate). In a previous report, it was suggested that the loss of the P450 function resulted in bacteria which were “shorter and fatter”, compared to wild type cells, but this thesis could find no evidence of such a phenomenon. CYP172A1 was successfully purified using recombinant expression in E. coli to enable biochemical and biophysical characterisation in vitro. CYP172A1 contains a typical P450 cysteine thiolate coordination to the heme iron, and exists in a low spin ferric heme state under neutral buffer conditions. The P450 was found to self aggregate, and despite rigorous investigations the cause of this aggregation was not fully established. Despite this issue, CYP172A1 was shown to bind to a wide range of P450 inhibitor-type compounds, with econazole displaying the tightest binding affinity (Kd = 100 nM). Identification of substrate-like compounds was achieved using high throughput compound screening, and a number of organic compounds were identified and shown to bind CYP172A1, inducing heme iron absorbance changes typical of either P450 inhibitors or substrates. Optical titrations for these molecules indicated that their CYP172A1 Kd values were in the low micromolar range. The catalytic capability of CYP172A1 was successfully demonstrated by providing the P450 with non native redox partners to oxidise one of such substrate-like compound (213071), resulting in the sulfoxidation of this compound.

Francisella tularensis is the causative agent of tularemia and is categorized by the CDC as a Tier 1 select agent. This gram-negative, facultative-intracellular bacterium infects macrophages by escaping the phagosome and replicating with high efficacy in the cytosol. Multiple virulence factors, including capsule and lipopolysaccharide (LPS), are expressed by F. tularensis. Biosynthesis of capsule and LPS O-antigen requires the same O-antigen biosynthesis gene cluster and, together, expression of capsule and O-antigen confer serum resistance. Mutations in the O-antigen biosynthesis gene cluster not only result in serum sensitivity, but also attenuate the ability to cause disease in vivo. In addition to changes in F. tularensis virulence, individual capsule and O-antigen mutants appear to have distinct intracellular phenotypes in macrophages. As previously shown by Lindemann et al. (2011), the capsule and O-antigen mutants FTT1236, FTT1237, and FTT1238 all replicated in human monocyte derived macrophages (MDMs) up to 16 hr and then ceased to replicate after that. This is hypothesized to be due to MDM cytotoxicity. In contrast, Raynaud et al. (2007) showed that the capsule and O-antigen mutant wbtA completely lacked replication in J774 macrophages, the reason for which has not been identified. A potential explanation for the loss of F. tularensis capsule and O-antigen mutant replication is capture and degradation by the host cell's autophagy pathway. Capture and degradation by autophagy is an accepted innate immune response to many intracellular pathogens. When small subpopulations of bacteria that normally replicate in membrane-bound vacuoles become cytosolic, such as Mycobacterium tuberculosis and Salmonella enterica serovar Typhimurium, they are targeted to forming autophagosomes through ubiquitination and binding of autophagy receptors. Pathogens have also developed methods to circumvent recognition and degradation by autophagy. Since F. tularensis replicates in the cytosol, it stands to reason that it has a means of evading detection by autophagy. We propose that expression of capsule and O-antigen acts as a mechanism used by F. tularensis to protect itself in an extracellular environment, as well as during intracellular infection. In this thesis we characterized nine different capsule and O-antigen mutants, and found different replication phenotypes in MDMs and varying degrees of MDM cytotoxicity. Also, only a subset of the mutants was detected by the autophagy marker, ubiquitin, supporting our hypothesis that different capsule and O-antigen mutants have diverse fates in MDMs. We also show that LVS and Schu S4 wbtA mutants had similar phenotypes. Upon further evaluation, we found that LVS wbtA more readily colocalized with ubiquitin, autophagy receptors, and the autophagy membrane protein LC3B, but not Beclin-1 or LAMP-1. This supports our hypothesis that capsule and O-antigen mutants are more susceptible to recognition by autophagy. Yet, because we did not observe LAMP-1 colocalization, there may be defects in the maturation of autophagosomes to degradative autolysosomes. Finally, we found that the fate of LVS wbtA in MDMs is dissimilar from J774 macrophages, suggesting macrophage species affect mutant fate. This thesis shows that different capsule and O-antigen mutants have multiple fates in MDMs, and suggest that F. tularensis capsule and O-antigen act as protective virulence factors that limit detection by autophagy.

No description available.

<p>During recent years, the field of computer vision has evolved rapidly. Convolutional Neural Networks (CNNs) have become the chosen default for implementing computer vision tasks. The popularity is based on how the CNNs have successfully performed the wellknown</p> <p>computer vision tasks such as image annotation, instance segmentation, and others with promising outcomes. However, CNNs have their caveats and need further research to turn them into reliable machine learning algorithms. The disadvantages of CNNs become more evident as the approach to breaking down an input image becomes apparent. Convolutional neural networks group blobs of pixels to identify objects in a given image. Such a</p> <p>technique makes CNNs incapable of breaking down the input images into sub-parts, which could distinguish the orientation and transformation of objects and their parts. The functions in a CNN are competent at learning only the shift-invariant features of the object in an image. The discussed limitations provides researchers and developers a purpose for further enhancing an effective algorithm for computer vision.</p> <p>The opportunity to improve is explored by several distinct approaches, each tackling a unique set of issues in the convolutional neural network’s architecture. The Capsule Network (CapsNet) which brings an innovative approach to resolve issues pertaining to affine transformations</p> <p>by sharing transformation matrices between the different levels of capsules. While, the Residual Network (ResNet) introduced skip connections which allows deeper networks</p> <p>to be more powerful and solves vanishing gradient problem.</p> <p>The motivation of these fusion of these advantageous ideas of CapsNet and ResNet with Squeeze and Excite (SE) Block from Squeeze and Excite Network, this research work presents SE-Residual Capsule Network (SE-RCN), an efficient neural network model. The proposed model, replaces the traditional convolutional layer of CapsNet with skip connections and SE Block to lower the complexity of the CapsNet. The performance of the model is demonstrated on the well known datasets like MNIST and CIFAR-10 and a substantial reduction in the number of training parameters is observed in comparison to similar neural networks. The proposed SE-RCN produces 6.37 Million parameters with an accuracy of 99.71% on the MNIST dataset and on CIFAR-10 dataset it produces 10.55 Million parameters with 83.86% accuracy.</p>

No / A computer simulation based on a Monte-Carlo technique has been developed and used to investigate the influence of pellet size, dispersity, shape and aggregation on the filling of hard shell capsules. The simulations are in general agreement with experimental observations previously reported. The results also confirm recent findings that filling is a function of pellet shape and that above an aspect ratio value of 1.2 filling reproducibility is reduced. The methodology is simple and rapid in execution allowing many computer-based experiments to be performed with minimum effort.

Streptococcus pneumoniae is a significant human pathogen and the leading cause of community-acquired pneumonia and acute otitis media. One of the primary defense mechanisms of the human immune system against pneumococcal infection involves granule-mediated killing of bacterial cells by neutrophils. While this mechanism has previously been shown to kill about half of pneumococci in vitro, we hypothesized that some pneumococcal strains have evolved to be more resistant to this granule-mediated killing. Clinical isolates demonstrated a varying range of sensitivity to neutrophil granules. Additionally, we established that the absence of the capsule may affect sensitivity as unencapsulated isolates showed a higher average survival than encapsulated isolates. Finally, pneumococcal surface protease HtrA was found to potentially serve as a protective factor as many knockouts were more sensitive than the wildtypes, recombinant HtrA protected wildtype TIGR4, and a resistant isolate showed higher htrA expression levels than sensitive isolates.

The effects caused to articular cartilage by the remote use of arthroscopically-delivered RF energy to soft tissues in the joint are unknown. This investigation reported the short and long-term effects of bRF and mRF energy on the biomechanical properties and metabolic activity of articular cartilage. In addition, the effect of Cosequin® therapy was addressed. Thirty New Zealand white rabbits were randomly assigned to one of two treatment groups (Group 1 - placebo; Group 2 - Cosequin®). Histopathology, cell viability, GAG synthesis, and mechanical function of the articular cartilage were compared between groups. Data were analyzed using a mixed model ANOVA (p=0.05). Immediate chondrocyte death was created by both RF devices. This damage was noted to be superficial and did not lead to the progressive deterioration of the extracellular matrix or mechanical function of the articular cartilage. Cosequin® therapy was unable to demonstrate significant differences compared to the control group.

Generative models are one of the most prominent components of unsupervised learning models that have a plethora of applications in various domains such as image-to-image translation, video prediction, and generating synthetic data where accessing real data is expensive, unethical, or compromising privacy. One of the main challenges in designing a generative model is creating a disentangled representation of generative factors which gives control over various characteristics of the generated data. Since the architecture of variational autoencoders is centered around latent variables and their objective function directly governs the generative factors, they are the perfect choice for creating a more disentangled representation. However, these architectures generate samples that are blurry and of lower quality compared to other state-of-the-art generative models such as generative adversarial networks. Thus, we attempt to increase the disentanglement of latent variables in variational autoencoders without compromising the generated image quality. In this thesis, a novel generative model based on capsule networks and a variational autoencoder is proposed. Motivated by the concept of capsule neural networks and their vectorized output, these structures are employed to create a disentangled representation of latent features in variational autoencoders. In particular, the proposed structure, called CapsuleVAE, utilizes a capsule encoder whose vector outputs can translate to latent variables in a meaningful way. It is shown that CapsuleVAE generates results that are sharper and more diverse based on FID score and a metric inspired by the inception score. Furthermore, two different methods for training CapsuleVAE are proposed, and the generated results are investigated. In the first method, an objective function with regularization is proposed, and the optimal regularization hyperparameter is derived. In the second method, called sequential optimization, a novel training technique for training CapsuleVAE is proposed and the results are compared to the first method. Moreover, a novel metric for measuring disentanglement in latent variables is introduced. Based on this metric, it is shown that the proposed CapsuleVAE creates more disentangled representations. In summary, our proposed generative model enhances the disentanglement of latent variables which contributes to the model's generalizing well to new tasks and more control over the generated data. Our model also increases the generated image quality which addresses a common disadvantage in variational autoencoders. / Master of Science / Generative models are algorithms that, given a large enough initial dataset, create data points (such as images) similar to the initial dataset from random input numbers. These algorithms have various applications in different fields, such as generating synthetic healthcare data, wireless systems data generation in extreme or rare conditions, generating high-resolution, colorful images from grey-scale photos or sketches, and in general, generating synthetic data for applications where obtaining real data is expensive, inaccessible, unethical, or compromising privacy. Some generative models create a representation for the data and divide it into several ``generative factors". Researchers have shown that a better data representation is one where the generative factors are ``disentangled", meaning that each generative factor is responsible for only one particular feature in the generated data. Unfortunately, creating a model with disentangled generative factors sacrifices the image quality. In this work, we design a generative model that enhances the disentanglement of generative factors without compromising the quality of the generated images. In order to design a generative model with more disentangled generative factors, we employ capsule networks in the architecture of the generative model. Capsule networks are algorithms that classify the inputted information into different categories. We show that by using capsule networks, our designed generative model achieves higher performance in the quality of the generated images and creates a more disentangled representation of generative factors.