Thesis (Ph. D.) -- California Institute of Technology, 2010. / Title from home page (viewed 05/13/10). Advisor and committee chair names found in the thesis' metadata record in the digital repository. Includes bibliographical references.

Thesis (M.S.E.C.E.)--University of Delaware, 2007. / Principal faculty advisors: Fouad Kiamilev and Robert F. Rogers, Dept. of Electrical and Computer Engineering. Includes bibliographical references.

Thesis (Ph. D.)--University of Texas at Austin, 2004. / Supervisor: Michael C. Downer. Vita. Includes bibliographical references.

Thesis (Ph.D)--Chemistry and Biochemistry, Georgia Institute of Technology, 2010. / Committee Chair: Dickson, Robert; Committee Member: Fahrni, Christoph; Committee Member: Payne, Christine; Committee Member: Petty, Jeff; Committee Member: Srinivasarao, Mohan. Part of the SMARTech Electronic Thesis and Dissertation Collection.

Thesis (Ph. D.)--University of Texas at Austin, 2001. / Vita. Includes bibliographical references. Available also from UMI Company.

Thesis (Ph. D.)--University of Texas at Austin, 2008. / Vita. Includes bibliographical references.

The unique symmetry requirements of second harmonic generation (SHG) provide exquisite selectivity to chiral crystals, enabling independent quantitative modeling of the nucleation and crystal growth of active pharmaceutical ingredients (APIs) within amorphous solid dispersions (ASDs) during accelerated in situ stability testing, and in vitro dissolution testing. ASDs, in which an API is maintained in an amorphous state within a polymer matrix, are finding increasing use to address solubility limitations of small-molecule APIs. SHG microscopy yielded limits of detection for ritonavir crystals as low as 10 ppm, which is about two orders of magnitude lower than other methods currently available for crystallinity detection in ASDs. The quantitative capabilities of SHG analysis were substantially improved further while simultaneously dramatically reducing the total sample volume and storage burden through in situ analysis. Single particle tracking of crystal growth performed in situ enabled substantial improvements in the signal to noise ratio (SNR) for recovered crystal nucleation and growth rates by nonlinear optical microscopy. Upon dissolution, the presence of solubilizing additives in biorelevant media greatly affected the generation and stabilization of supersaturated solutions. SHG microscopy was found to enable the detection of crystals even in the highly turbid Ensure Plus® system. Analysis of the SHG micrographs clearly indicated that differences in the nucleation kinetics rather than growth rates dominated the overall trends in crystallinity. For weakly basic drugs, the fate of dissolution in fasted-state simulated intestinal fluid (FaSSIF, pH 6.5) varied with the ASDs drug loading, and was highly affected by the pre-exposure to the fasted-state simulated gastric fluid (FaSSGF, pH 1.6) medium, versus the dissolution in FaSSIF medium alone. The presence of crystals during the first stage of posaconazole ASDs dissolution in FaSSGF acted as nuclei for further crystallization in the later dissolution stage in FaSSIF. The results provide insights of better formulation prediction of poorly soluble drugs, as well as understanding origins of intraluminal absorption variability for such systems

The two main themes in this thesis are the implementation of interference methods with optically trapped particles for measurements of position and optical phase (optical interferometric metrology) and the optical manipulation of nanoparticles for studies in the assembly of nanostructures, nanoscale heating and nonlinear optics (nanophotonics). The first part of the thesis (chapter 1, 2) provides an introductory overview to optical trapping and describes the basic experimental instrument used in the thesis respectively. The second part of the thesis (chapters 3 to 5) investigates the use of optical interferometric patterns of the diffracting light fields from optically trapped microparticles for three types of measurements: calibrating particle positions in an optical trap, determining the stiffness of an optical trap and measuring the change in phase or coherence of a given light field. The third part of the thesis (chapters 6 to 8) studies the interactions between optical traps and nanoparticles in three separate experiments: the optical manipulation of dielectric enhanced semiconductor nanoparticles, heating of optically trapped gold nanoparticles and collective optical response from an ensemble of optically trapped dielectric nanoparticles.

<div>Fluorescence microscopy has emerged as a powerful tool for studying cell biology because it enables the acquisition of 3D image volumes deeper into tissue and the imaging of complex subcellular structures. Fluorescence microscopy images are frequently distorted by motion resulting from animal respiration and heartbeat which complicates the quantitative analysis of biological structures needed to characterize the structure and constituency of tissue volumes. This thesis describes a two pronged approach to quantitative analysis consisting of non-rigid registration and deep convolutional neural network segmentation. The proposed image registration method is capable of correcting motion artifacts in three dimensional fluorescence microscopy images collected over time. In particular, our method uses 3D B-Spline based nonrigid registration using a coarse-to-fine strategy to register stacks of images collected at different time intervals and 4D rigid registration to register 3D volumes over time. The results show that the proposed method has the ability of correcting global motion artifacts of sample tissues in four dimensional space, thereby revealing the motility of individual cells in the tissue.</div><div><br></div><div>We describe in thesis nuclei segmentation methods using deep convolutional neural networks, data augmentation to generate training images of different shapes and contrasts, a refinement process combining segmentation results of horizontal, frontal, and sagittal planes in a volume, and a watershed technique to enumerate the nuclei. Our results indicate that compared to 3D ground truth data, our method can successfully segment and count 3D nuclei. Furthermore, a microscopy image synthesis method based on spatially constrained cycle-consistent adversarial networks is used to efficiently generate training data. A 3D modified U-Net network is trained with a combination of Dice loss and binary cross entropy metrics to achieve accurate nuclei segmentation. A multi-task U-Net is utilized to resolve overlapping nuclei. This method was found to achieve high accuracy object-based and voxel-based evaluations.</div>

Maîtriser l’impact des maladies sur les cultures est un défi majeur de l’agriculture moderne. Cette préoccupation est un aspect important de l’optimisation de la productivité, notamment en viticulture. En France, le mildiou de la vigne causé par Plasmopara viticola est une des maladies cryptogamiques responsable des épidémies les plus dévastatrices et les plus redoutées. Les traitements reposent sur l’utilisation préventive, systématique et onéreuse de composés chimiques antifongiques dont l’utilisation massive constitue un risque à la fois pour l’homme et l’environnement. La réduction de l’utilisation de fongicide implique le développement d’outils de diagnostic au champ, qui requiert la compréhension des interactions entre la plante et les agents pathogènes. Les travaux de cette thèse pluridisciplinaire ont porté sur le pathosystème Plasmopara viticola - Vitis vinifera, notamment pour répondre à l’intérêt croissant pour un outil de diagnostic en temps réel de la maladie utilisable au vignoble. Les stilbènes sont des phytoalexines impliqués dans la défense de certaines plantes supérieures vis-à-vis de stress biotiques et abiotiques. L’autofluorescence de ces composés phénoliques, dont la biosynthèse est induite dans les feuilles de vigne par P. viticola, en fait un potentiel marqueur naturel de l’infection. En effet, la faible autofluorescence bleu-verte des feuilles de vigne saines est considérablement renforcée par l’autofluorescence violet-bleue des stilbènes à la surface de feuilles de vigne infectée par P. viticola. Cette étude a montré que quelque soit le niveau de résistance du génotype, l’autofluorescence violet-bleue des stilbènes induit par l’infection est présente au niveau des parois des cellules de l’épiderme. En dehors de la concentration, la viscosité s’est révélé être la principale variable physico-chimique influençant l’intensité de l’autofluorescence des stilbènes dans les différents compartiments cellulaires des feuilles de vigne. Ceci explique la fluorescence intense des parois, particulièrement rigides, des cellules de garde (stomates) des feuilles infectées. Le suivi cinétique journalier a révélé la nature transitoire de l’autofluorescence des stilbènes lors de l’infection. La robustesse et l’intérêt de ce signal a également été validée par la mesure à différentes échelles (de la cellule à la feuille entière) et avec différentes méthodes fluorimétriques. Les résultats de ce travail ont permis des avancées sur la connaissance du rôle de composés phénoliques induits et constitutifs dans la défense contre P. viticola. En plus de la localisation de l’autofluorescence des stilbènes en surface des feuilles, la microscopie confocale couplée à la microspectrofluorimetrie a révélé différentes localisations de ces phytoalexines dans la profondeur des tissus en corrélation avec le niveau de résistance des génotypes. L’utilisation de l’autofluorescence des stilbènes comme marqueur de l’infection a permis de mettre en évidence : 1) le fait que les flavonols constitutifs des feuilles de V. vinifera retardent le développement de l’infection par P. viticola; et 2) le fait que les acides hydroxycinnamiques constitutifs ne semble pas participer à la défense contre P. viticola. Enfin, une nouvelle méthode de diagnostic non-destructive du mildiou sur feuille basée sur l’autofluorescence des stilbènes a été développée. Elle a montré une détection pré-symptomatique du mildiou sur les feuilles de vigne entières dès le premier jour après l’infection sur la face abaxiale et le troisième jour sur la face adaxiale. Cette méthode de diagnostic du mildiou a été validée au laboratoire notamment grâce à un prototype de capteur proximal développé en collaboration avec la société Force-A. La validation de la méthode au vignoble dans le cadre d’infection naturelle est la prochaine étape pour une utilisation de ce capteur optique dans le cadre de l’agriculture durable et de la sélection variétale. / Controlling the impact of diseases on crops is a major challenge of modern agriculture. This concern is an important aspect of optimizing productivity, notably in viticulture. In France, downy mildew caused by Plasmopara viticola is a fungal disease responsible for the most devastating epidemics. The preventive and systematic treatments are expensive, while the massive use of antifungal chemicals is a risk to both humans and the environment. Reducing the use of fungicide involves the development of diagnostic tools in the field, which requires understanding the interactions between plants and pathogens. The work of this multidisciplinary thesis focused on the pathosystem Plasmopara viticola - Vitis vinifera, especially to meet the growing interest in a real-time diagnostic tool of disease applicable in the vineyard. Stilbenes are phytoalexins involved in the defense of certain higher plants against biotic and abiotic stresses. The autofluorescence of these phenolic compounds, whose biosynthesis is induced in grapevine leaves by P. viticola, makes it a potential marker of natural infection. Indeed, the low blue-green autofluorescence of grapevine leaves is greatly enhanced by the violet-blue autofluorescence of stilbenes on the surface of leaves infected by P. viticola. This study showed that whatever the level of resistance in various genotypes, violet-blue autofluorescence induced by stilbene is present in the walls of epidermal cells. In addition to their concentration, viscosity proved the main physico-chemical variable affecting the intensity of the autofluorescence of stilbenes in different compartments of vine leaves. This explains the intense fluorescence of the walls, particularly rigid, of guard cells (stomata) of infected leaves. Daily monitoring revealed a kinetic with a transient rise of the autofluorescence of stilbenes during infection. The robustness and value of this signal was also validated by measuring at different levels (cellular to whole leaf) and with various fluorimetric methods (imaging, spectroscopy, proximal sensing). These results advance our understanding of the role of constitutive and induced phenolic compounds in plant defence against P. viticola. In addition to a common location of the autofluorescence of stilbenes on the leaf surface, confocal microscopy coupled with microspectrofluorometry revealed distinctive localizations of these phytoalexins in the deep tissue correlated with the level of resistance in genotypes. This aspect no doubt needs broader testing. The use of autofluorescence of stilbene as a marker of infection allowed us to ascertain that: 1) constitutive flavonols of the leaves of V. vinifera retard the development of infection by P. viticola and 2) the constitutive hydroxycinnamic acids do not seem to participate in the defence against P. viticola. Finally, a new method for the non-destructive diagnosis of leaf infection based on the autofluorescence of stilbenes has been developed. We have demonstrated a pre-symptomatic detection of downy mildew on whole grape leaves from the first day after infection on the abaxial surface and from the third day on the adaxial surface. This method of diagnosis has been validated in the laboratory thanks to a proximal sensor prototype developed in collaboration with the company Force-A. The validation of the method in the vineyard in a context of natural infections is the next step for use of this optical sensor as a tool for sustainable agriculture and for genetic screening.