Knowledge of the mechanics of the lens capsule is crucial for improving cataract surgery as well as understanding better the physiological role of the lens capsule in the process of accommodation. Previous research on the mechanical properties of the lens capsule contains many gaps and contradictions due to experimental limitations and inappropriate assumptions. Thus, the goal of this work is to quantify fully the regional, multiaxial mechanical behavior of the lens capsule and to calculate the change in stress and strain fields as a result of cataract surgery. Determining in situ the multiaxial mechanical behavior of the lens capsule required the design and construction of an experimental device capable of altering stresses in the capsule while measuring localized surface deformations. Tests performed on this device reveal that the meridional and circumferential strains align with the principal directions and are equivalent through most of the anterior lens capsule, except close to the equator where the meridional strain is greater. Furthermore, preconditioning effects were also found to be significant. Most importantly, however, these tests provide the data necessary for calculating material properties. This experimental system is advantageous in that it allows reconstruction of 3D geometry of the lens capsule and thereby quantification of curvature changes, as well as measurement of surface deformations that result from various surgical interventions. For instance, a continuous circular capsulorhexis (CCC) is commonly used during cataract surgery to create a hole in the anterior lens capsule (typically with a diameter of 5 mm). After the introduction of a CCC, strain was found to redistribute evenly from the meridional direction (retractional strain) to the circumferential direction (extensional strain), where both directional components of strain reached magnitudes up to 20% near the edge of the CCC. Furthermore, the curvature was found to increase at the edge of the CCC and remain the same near the equator, indicating that the mere introduction of a hole in the lens capsule will alter the focal characteristics of the lens and must therefore be considered in the design of an accommodative intraocular lens.

Indiana University-Purdue University Indianapolis (IUPUI) / Machine learning models are rising every day. Most of the Computer Vision oriented machine learning models arise from Convolutional Neural Network’s(CNN) basic structure. Machine learning developers use CNNs extensively in Image classification, Object Recognition, and Image segmentation. Although CNN produces highly compatible models with superior accuracy, they have their disadvantages. Estimating pose and transformation for computer vision applications is a difficult task for CNN. The CNN’s functions are capable of learning only shift-invariant features of an image. These limitations give machine learning developers motivation towards generating more complex algorithms. Search for new machine learning models led to Capsule Networks. This Capsule Network was able to estimate objects’ pose in an image and recognize transformations to these objects. Handwritten digit classification is the task for which capsule networks are to solve at the initial stages. Capsule Networks outperforms all models for the MNIST dataset for handwritten digits, but to use Capsule networks for image classification is not a straightforward multiplication of parameters. By replacing the Capsule Network’s initial layer, a simple Convolutional Layer, with complex architectures in CNNs, authors of Residual Capsule Network achieved a tremendous change in capsule network applications without a high number of parameters. This thesis focuses on improving this recent Residual Capsule Network (RCN) to an extent where accuracy and model size is optimal for the Image classification task with a benchmark of the CIFAR-10 dataset. Our search for an exemplary capsule network led to the invention of RCN2: Residual Capsule Network 2 and RCNX: Residual Capsule NeXt. RCNX, as the next generation of RCN. They outperform existing architectures in the domain of Capsule networks, focusing on image classification such as 3-level RCN, DCNet, DC Net++, Capsule Network, and even outperforms compact CNNs like MobileNet V3. RCN2 achieved an accuracy of 85.12% with 1.95 Million parameters, and RCNX achieved 89.31% accuracy with 1.58 Million parameters on the CIFAR-10 benchmark.

The pathogenic actinomycete Rhodococcus equi is a facultative intracellular parasite that replicates within macrophages. This ability is dependent on the pVAP virulence plasmid, and more specifically, on the laterally acquired vap pathogenicity island (vap PAI) carried by it. R. equi has two contrasting lifestyles as a soil-dwelling microbe and as an inhabitant of the intracellular macrophage compartment. In the first part of this thesis work we analysed the nature of the signals recognised by R. equi to adapt the expression of the virulence genes of the plasmid during the transition from soil saprotroph to intracellular parasite. The expression profile of virulence plasmid genes in response to temperature and pH in vitro and to the macrophage environment was investigated by microarray analysis. A shift to 37ºC was the main stimulus involved in vap PAI gene activation and macrophage-derived signals did not further modulate the expression of the PAI genes contrary to previous suggestions. In a second part of the thesis we investigated the role of a horizontally acquired island encoding exopolysaccharide biosynthesis in the R. equi saprotroph-intracellular parasite dual lifestyle. Mutational analysis of this locus showed that it is responsible for the typical mucoid colony morphology of R. equi and the ability to produce a polysaccharide capsule. Mutations in the capsule locus favoured macrophage uptake but had no effect on intracellular proliferation and in vivo survival in mice. However, the capsule mutants showed significantly increased susceptibility to desiccation, ultraviolet radiation and heat and were outcompeted by capsulated wild-type R. equi in dry soil. Thus, while having a minor role in virulence, the R. equi capsule appears to be primarily required for survival in soil and to act as a transmission factor. The third part of this work followed the identification of a horizontally acquired locus that encodes pili appendages that promote association with macrophages and colonization of the mouse lung. The ability of a component of this structure, the RplB pilin subunit, to act as a vaccine antigen was investigated in mice and horses. Vaccinated mice produced high levels of anti-RplB IgG and showed significant protection against pulmonary challenge with virulent R. equi. The experimental RplB subunit vaccine proved also to be immunogenic in horses, eliciting a strong IgG response in pregnant mares and foals. We also demonstrated passive transfer of high levels of maternal anti-RplB antibodies from the mares to the foals via colostrum. Our results indicate that the RplB pilin subunit is a promising novel candidate R. equi vaccine antigen.

Francisella tularensis is a Gram-negative pathogenic organism that causes the disease tularemia. This disease can be potentially fatal without treatment. Francisella tularensis virulent strains can cause disease in humans with an infectious dose as low as 10 organisms. As a result of this low infectious dose, high mortality, and ease to produce an aerosol inoculum, the Centers for Disease Control and Prevention has classified Francisella tularensis as a Tier I select agent, the highest threat level. Much research has been done to determine the cause for the extreme virulence. However, despite these efforts, little is known about the mechanisms by which Francisella goes undetected inside host cells until it is too late for the host to respond. Researchers in the Jones' laboratory utilized a transposon site hybridization (TraSH) screen with human monocyte derived macrophages (MDMs) as the host cell and an enzyme-linked immunosorbent assay (ELISA) screen of pools of transposon mutants searching for virulence determinants and genes responsible for Francisella capsule or LPS. Through the TraSH screen, our group identified a locus of genes, FTT1236, FTT1237, and FTT1238c as being important for survival within human MDMs. From the mutant library screen using ELISA, I identified the same genes, FTT1236 and FTT1238c. In addition, I also identified wzy, wbtA, FTT0846, and hemH as being involved in LPS and or capsule production. A similar ELISA screen was done by researchers in Apicella laboratory using a different monoclonal antibody that identified insertions in, dnaJ, manB and an intergenic region between FTT0673 and FTT0674c that potentially disrupted LPS and capsule biogenesis. Previously, FTT1236, FTT1237, and FTT1238c mutants were observed by our laboratory to be serum sensitive and activate MDMs by an unknown mechanism. I further characterized these mutant strains by analyzing the changes in the LPS core. I identified core truncations for the FTT1236 and FTT1237 mutants, but not FTT1238c. Combining this new data with previously published work and bioinformatical analysis of the FTT1236, FTT1237 and FTT1238c proteins, I hypothesized that these proteins have functions similar to Waa proteins of other organisms, which are involved in LPS core assembly and O-antigen ligation. With functional complementation and mass spectrometry of LPS preparations, I have designated FTT1236, FTT1237, and FTT1238c as WaaY, WaaZ, and WaaL respectively. In addition to this work characterizing the biochemical functions of these gene products, I examined the effect of mutations in these genes on the virulence of Francisella. In contrast to infection with wild type Schu S4, mice infected either intraperitoneally or intranasally displayed significant inflammatory responses to infection and the strains were significantly attenuated by either route of infection. I also observed that waaY and waaL mutant strains disseminated to the liver and spleen after an intranasal infection despite their lack of O-antigen and capsule. At an i.n. dose of 106 CFU these mutant strains still caused lethal murine infection, but death occurred around day 12 post infection; mice infected with <20 CFU of Schu S4 succumb at day 5 post infection. The cause of the death in mice infected with these mutant strains was pulmonary edema, rather than multiple organ failure induced by Schu S4. Of the additional seven mutant strains identified from the ELISA screens, I characterized their physical phenotypes, virulence defects, and their potential as an attenuated live vaccine. All of these strains were determined to be sensitive to human pooled serum to various degrees. Three of these strains, dnaJ::Tn5, hemH::Tn5, and FTT0673p/prsAp::Tn5 did not have identifiable defects in capsule or LPS biosynthesis, nor were they attenuated in mice. The remaining four strains, FTT0846::Tn5, manB::Tn5, wzy::Tn5, and wbtA::Tn5, were found to have LPS O-antigen and capsule defects, and two of these strains had LPS core defects (FTT0846::Tn5 and manB::Tn5). Each of these four strains was attenuated in mice, when compared to WT. I also tested the ability of mice infected with waaY::TrgTn, waaL::TrgTn, and wbt::Tn5 to be protected from lethal challenges of Schu S4. All three strains provided some level of protection against lethal Schu S4 challenges. In addition, I also tested Francisella LPS and capsule to provide protection against lethal challenges of LVS and Schu S4. I determined that LPS and capsule protected against high doses of LVS, but LPS did not provide any protection when immunized mice were challenged with Schu S4. Interestingly, we observed that mice immunized with capsule were partially protected from lethal Schu S4 challenges. In addition, I observed a novel difference between virulent Francisella strains and LVS, in that virulent strains have O-antigen glycosylated and LVS appears to be lacking this characteristic. Collectively, this work adds to the growing data of the importance of LPS and the role of capsule role in immune evasion as well as the significance of capsule and LPS mutant strains to provide protection against Schu S4.

Streptococcus pyogenes is a re-emerging pathogen that produces superficial and life threatening invasive diseases. One important virulence factor in S. pyogenes is hyaluronic acid capsule which has been shown to increase expression at sub-body temperatures in certain strains. This study showed that thermoregulation is common in invasive clinical isolates. Regulation was shown to occur independent of the CovRS two-component regulator in a post transcription manner and before protein level regulation. The endoribonuclease, CvfA, was also confirmed to be required for capsule thermoregulation. The search for a regulator lead to the discovery of the 1st antisense RNAs in S. pyogenes found opposite the capsule synthesis genes. Its role if any in capsule has not been discovered. Finally, a group of sRNAs were characterized adding to the knowledge of this layer of regulation in S. pyogenes.

Capsule Networks are the next generation of image classifiers. Although they have several advantages over conventional Convolutional Neural Networks (CNNs), they remain computationally heavy. Since inference on Capsule Networks is timeconsuming, thier usage becomes limited to tasks in which latency is not essential. Approximation methods in Deep Learning help networks lose redundant parameters to increase speed and lower energy consumption. In the first part of this work, we go through an algorithm called zero-skipping. More than 50% of trained CNNs consist of zeros or values small enough to be considered zero. Since multiplication by zero is a trivial operation, the zero-skipping algorithm can play a massive role in speed increase throughout the network. We investigate the eligibility of Capsule Networks for this algorithm on two different datasets. Our results suggest that Capsule Networks contain enough zeros in their Primary Capsules to benefit from this algorithm. In the second part of this thesis, we investigate pruning as one of the most popular Neural Network approximation methods. Pruning is the act of finding and removing neurons which have low or no impact on the output. We run experiments on four different datasets. Pruning Capsule Networks results in the loss of redundant Primary Capsules. The results show a significant increase in speed with a minimal drop in accuracy. We also, discuss how dataset complexity affects the pruning strategy. / Graduate

Doctor of Philosophy / Department of Chemistry / Stefan Kraft / In the present study, the rescorcin[4]arenes decorated with an alternating arrangement of pyridinium and sulfonate moieties were synthesized to establish a molecular quadrupole. These rescorcin[4]arenes are capable of undergoing conformational changes from a ‘collapsed’ to ‘open’ state. The stability of the two state are controlled by the polarity of the solvent environment. The electrostatic interactions between pyridinium and sulfonates enforced a ‘quadrupolar collapse’ in solvents such as chloroform. While these interactions are disrupted in dimethylsulfoxide. A major synthetic challenge was functionalizing the resorcin[4]arene in a positive/negative/positve/negative pattern was successfully addressed. Comparison to dipolar resorcin[4]arene were undertaken to establish a quantitative measurement of the quadrupolar forces and to address the question of cooperatively provided addition attraction beyond two dipoles. A,C-functionalized-bis sulfonate resorcin[4]arenes and A,C function bis-pyridinium resocin[4]arenes were synthesized independently. Combining these dicationic and dianionic moieties provided an interdigitating dimeric unit with overall quadrupolar charge distribution. Disruption of the quadrupolar salt bridges in CDCl[subscript]3 was accomplished through the addition of DMSO or through dilution. DeltaG[superscript]o, DeltaH[superscript]o, and DeltaS[superscript]o have been determined for the dimer formation. Addition of pyridinium salts led to a disruption of the dimeric capsule. Host-guest binding studies established attractive binding to CS[superscript]2. Larger guests such as toluene, diiodobenzene, dicynanobenzene could not be encapsulated.

Cet ouvrage porte sur la conception, la synthèse et la caractérisation de sphères supramoléculaires auto-assemblées. Si le groupe Dory a, par le passé, été en mesure d’effectuer la synthèse d’un type de sphères avec 4 unités monomères, le but de cette recherche a été de développer des monomères qui tendraient vers des sphères à 8 unités voire même 20. Suite à de nombreuses idées, nous avons orienté notre recherche davantage sur un type de monomère afin de pouvoir le synthétiser en quantité suffisamment importante pour pouvoir le caractériser. Une idée aussi novatrice a nécessité plusieurs mois de travail pour parvenir à un produit satisfaisant et suffisamment pur pour commencer la caractérisation.

No description available.

Francisella tularensis was a facultative intracellular pathogen and a gram-negative coccobacillus which has been categorized by the CDC as a potential class A select agent due to its highly infectious properties and high mortality rates. Francisella tularensis was also responsible for the zoonotic disease tularemia, which was usually transmitted by arthropod vectors or via contact with infected animals. Francisella tularensis subspecies novicida has been used by many researchers in genetic pathogenesis experiments to try to elucidate genes responsible for virulence factors. One of these virulence factors was a capsular material which has been thought to be involved in either increasing pathogenicity or infectivity of this organism upon engulfment by its principal host cell, the macrophage. There were many potential genetic loci which may be involved in this biosynthetic process of encapsulation. One such locus has excellent homology to the capsule biosynthesis operon of Bacillus anthracis, which, under certain conditions, creates a polyglutamic acid capsule (PGA). A transposon mutation in the amide ligase (capB) in LVS has a reduced virulence in murine infection models. I wished to investigate whether Francisella novicida was capable of producing such a capsule and under which environmental conditions this capsule was made. I have created a site-directed mutant of the capB gene in Francisella novicida U112 using targeted mutagenesis via PCR SOEing and have introduced this mutation via electroporation of a suicide vector. I have tested our mutant against preimmune serum treatments and have shown reduced viability as well as a reduced capacity for replication inside RAW 264.7 murine macrophages. I assayed for production of a PGA capsule via immunodot blot and electron microscopy as well as analysis by mass spectrophotometry of capsular extracts. I also tested various media constituents and different environmental conditions to determine which external stimuli may contribute to PGA capsule biosynthesis as well as regulatory changes in transcript levels of this operon.