Dissertations / Theses on the topic 'From-To matrix'

Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro de comunicação e Expressão, Programa de Pós-Graduação em Letras/Inglês e Literatura Correspondente.<br>Made available in DSpace on 2012-10-20T02:49:28Z (GMT). No. of bitstreams: 0Bitstream added on 2014-09-26T01:01:56Z : No. of bitstreams: 1 182188.pdf: 2777855 bytes, checksum: 7b4d6a42680bc77daf11ab9ee93f46bd (MD5)<br>Este trabalho lida com o gênero literário de ficção científica. Combinando os princípios da "Cultural Criticism" e "Reader-Response Criticism," ele discute e interpreta duas narrativas ocidentais: Frankenstein de Mary Shelley e Matrix dos irmãos Wachowski. O trabalho se preocupa com dois aspectos relevantes: (1) a cultura Ocidental esmagando dependência na ciência e tecnologia e (2) o papel das narrativas como um instrumento de ambos apoio e mudança em relação aos valores e verdades propostas pelo discurso dominante ou cultura paradigmática.

De nos jours, l'IA repose en grande partie sur l'utilisation de données de grande taille et sur des méthodes d'apprentissage machine améliorées qui consistent à développer des algorithmes de classification et d'inférence en tirant parti de grands ensembles de données de grande taille. Ces grandes dimensions induisent de nombreux phénomènes contre-intuitifs, conduisant généralement à une mauvaise compréhension du comportement de nombreux algorithmes d'apprentissage machine souvent conçus avec des intuitions de petites dimensions de données. En tirant parti du cadre multidimensionnel (plutôt que d'en souffrir), la théorie des matrices aléatoires (RMT) est capable de prédire les performances de nombreux algorithmes non linéaires aussi complexes que certains réseaux de neurones aléatoires, ainsi que de nombreuses méthodes du noyau telles que les SVM, la classification semi-supervisée, l'analyse en composantes principales ou le regroupement spectral. Pour caractériser théoriquement les performances de ces algorithmes, le modèle de données sous-jacent est souvent un modèle de mélange gaussien (MMG) qui semble être une hypothèse forte étant donné la structure complexe des données réelles (par exemple, des images). En outre, la performance des algorithmes d'apprentissage automatique dépend du choix de la représentation des données (ou des caractéristiques) sur lesquelles ils sont appliqués. Encore une fois, considérer les représentations de données comme des vecteurs gaussiens semble être une hypothèse assez restrictive. S'appuyant sur la théorie des matrices aléatoires, cette thèse vise à aller au-delà de la simple hypothèse du MMG, en étudiant les outils classiques d'apprentissage machine sous l'hypothèse de vecteurs aléatoires concentrés qui généralisent les vecteurs Gaussiens. Cette hypothèse est particulièrement motivée par l'observation que l'on peut utiliser des modèles génératifs (par exemple, les GAN) pour concevoir des structures de données complexes et réalistes telles que des images, grâce à des transformations Lipschitzienne de vecteurs gaussiens. Cela suggère notamment que l'hypothèse de concentration sur les données mentionnée ci-dessus est un modèle approprié pour les données réelles et qui est tout aussi mathématiquement accessible que les MMG. Par conséquent, nous démontrons à travers cette thèse, en nous appuyant sur les GANs, l'intérêt de considérer le cadre des vecteurs concentrés comme un modèle pour les données réelles. En particulier, nous étudions le comportement des matrices de Gram aléatoires qui apparaissent au cœur de divers modèles linéaires, des matrices à noyau qui apparaissent dans les méthodes à noyau et également des méthodes de classification qui reposent sur une solution implicite (par exemple, la couche de Softmax dans les réseaux de neurones), avec des données aléatoires supposées concentrées. En particulier, la compréhension du comportement de ces matrices/méthodes, pour des données concentrées, nous permet de caractériser les performances (sur des données réelles si nous les assimilons à des vecteurs concentrés) de nombreux algorithmes d'apprentissage machine, tels que le clustering spectral, les SVM, l'analyse en composantes principales et l'apprentissage par transfert. L'analyse de ces méthodes pour des données concentrées donne le résultat surprenant qu'elles ont asymptotiquement le même comportement que pour les données de MMG. Ce résultat suggère fortement l'aspect d'universalité des grands classificateurs d'apprentissage machine par rapport à la distribution sous-jacente des données<br>AI nowadays relies largely on using large data and enhanced machine learning methods which consist in developing classification and inference algorithms leveraging large datasets of large sizes. These large dimensions induce many counter-intuitive phenomena, leading generally to a misunderstanding of the behavior of many machine learning algorithms often designed with small data dimension intuitions. By taking advantage of (rather than suffering from) the multidimensional setting, random matrix theory (RMT) is able to predict the performance of many non-linear algorithms as complex as some random neural networks as well as many kernel methods such as Support Vector Machines, semi-supervised classification, principal component analysis or spectral clustering. To characterize the performance of these algorithms theoretically, the underlying data model is often a Gaussian mixture model (GMM) which seems to be a strong assumption given the complex structure of real data (e.g., images). Furthermore, the performance of machine learning algorithms depends on the choice of data representation (or features) on which they are applied. Once again, considering data representations as Gaussian vectors seems to be quite a restrictive assumption. Relying on random matrix theory, this thesis aims at going beyond the simple GMM hypothesis, by studying classical machine learning tools under the hypothesis of Lipschitz-ally transformed Gaussian vectors also called concentrated random vectors, and which are more generic than Gaussian vectors. This hypothesis is particularly motivated by the observation that one can use generative models (e.g., GANs) to design complex and realistic data structures such as images, through Lipschitz-ally transformed Gaussian vectors. This notably suggests that making the aforementioned concentration assumption on data is a suitable model for real data and which is just as mathematically accessible as GMM models. Moreover, in terms of data representation, the concentration framework is compatible with one of the most widely used data representations in practice, namely deep neural nets (DNNs) representations, since they consist in a Lipschitz transformation of the input data (e.g., images). Therefore, we demonstrate through this thesis, leveraging on GANs, the interest of considering the framework of concentrated vectors as a model for real data. In particular, we study the behavior of random Gram matrices which appear at the core of various linear models, kernel matrices which appear in kernel methods and also classification methods which rely on an implicit solution (e.g., Softmax layer in neural networks), with concentrated random inputs. Indeed, these methods are at the heart of many classification, regression and clustering machine learning algorithms. In particular, understanding the behavior of these matrices/methods, for concentrated data, allows us to characterize the performances (on real data if we assimilate them to concentrated vectors) of many machine learning algorithms, such as spectral clustering, SVMs, principal component analysis and transfer learning. Analyzing these methods for concentrated data yields to the surprising result that they have asymptotically the same behavior as for GMM data (with the same first and second order statistics). This result strongly suggest the universality aspect of large machine learning classifiers w.r.t. the underlying data distribution

Sixty years after the works of Wigner and Dyson, Random Matrix Theory still remains a very active and challenging area of research, with countless applications in mathematical physics, statistical mechanics and beyond. In this thesis, we focus on rotationally invariant models where the requirement of independence of matrix elements is dropped. Some classical examples are the Jacobi and Wishart-Laguerre (or chiral) ensembles, which constitute the core of the present work. The Wishart-Laguerre ensemble contains covariance matrices of random data, and represents a very important tool in multivariate data analysis, with recent applications to finance and telecommunications. We will first consider large deviations of the maximum eigenvalue, providing new analytical results for its large N behavior, and then a power-law deformation of the classical Wishart-Laguerre ensemble, with possible applications to covariance matrices of financial data. For the Jacobi matrices, which arise naturally in the quantum conductance problem, we provide analytical formulas for quantities of interest for the experiments.

Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2016.<br>This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.<br>Cataloged from student-submitted PDF version of thesis.<br>Includes bibliographical references (pages 37-39).<br>Neighborhood-based Collaborative filtering (CF) methods have proven to be successful in practice and are widely applied in commercial recommendation systems. Yet theoretical understanding of their performance is lacking. In this work, we introduce a new framework of Blind Regression which assumes that there are latent features associated with input variables, and we observe outputs of some Lipschitz continuous function over those unobserved features. We apply our framework to the problem of matrix completion and give a nonparametric method which, similar to CF, combines the local estimates according to the distance between the neighbors. We use the sample variance of the difference in ratings between neighbors as the proximity of the distance. Through error analysis, we show that the minimum sample variance is a good proxy of the prediction error in the estimates. Experiments on real-world datasets suggests that our matrix completion algorithm outperforms classic user-user and item-item CF approaches. Finally, our framework easily extends to the setting of higher-order tensors and we present our algorithm for tensor completion. The result from real-world application of image inpainting demonstrates that our method is competitive with the state-of-the-art tensor factorization approaches in terms of predictive performance.<br>by Yihua Li.<br>M. Eng.

<p> Mechanical signals affect virtually every fundamental single- and multi-cellular process in biology. The local responses of individual molecules to mechanical stimuli at the interface of cell with its adjacent microenvironment (extracellular matrix or material) elicit global responses at the cell and tissue scales. Understanding and manipulating the cell-material interaction can be leveraged to design biomaterials with unique characteristics tailored towards a wide variety of biological applications such as platforms that direct stem cell differentiation for tissue engineering, sensors that can record accurate electrical signals in single cells for neuroscience, and implants that are susceptible to cell adhesion for biomedical applications. In this thesis I present work characterizing the response of cells to mechanical stimuli at the single cell and single molecule scales. At the single cell scale, we provide insights into how mechanical signals such as micro- and nano-topography of metallic and metallic surfaces affect cell adhesion, both in mammalian and bacterial cells. Next we characterize the mechanical response of protein complexes involved in the transmission of mechanical signals across the cytoskeleton to the nucleus. </p><p> The four main contributions of the work presented in this thesis are as follows: 1) We used high resolution scanning electron microscopy to characterize the cell-nanostructure interface and provide insights into the response of individual mammalian cells to nanostructures with complex geometries. 2) We provide a first look at how individual bacterial cells adhere to metallic nanostructures, which could lead to new techniques to thwart infections. 3) We proposed a novel technique to control the growth and arrangements of bacterial cell communities. This method will allow precise small-scale mechanical manipulation of bacterial cells and could be utilized for unraveling the understudied mechanisms of bacterial mechanosensitivity. 4) We performed the first molecular dynamics study on the mechanisms of force transmission to the nucleus of eukaryotic cells through protein complexes known as linkers of the nucleoskeleton and cytoskeleton (LINC complexes). We showed that LINC complexes are highly stable under tensile forces, and that the transmission of force across the complex depends highly on the unique intermolecular covalent bonds formed between the two proteins that construct the complex. Finally, we presented a model for the molecular mechanisms of LINC complex activation and regulation at the nuclear envelope.</p><p>

<p> Coordinated movement of cardiac valves controls unidirectional flow of the blood with every heart beat. Cardiac valves are composed of thin, pliable leaflets that withstand compressive tension, fluid shear stress, and bending stress as blood flows through them. The structure and the mechanical properties of the valves render them durable during the lifetime of human beings. However, changes in hemodynamic environment, inflammatory responses, and congenital valvular defects can all cause valves to undergo irreversible structural changes, one of which is calcific aortic stenosis (CAS). CAS affects 2-3% of the population over 65 years old in the western world, and the only effective treatment is valve replacement surgery. CAS is characterized by tissue stiffening and the formation of calcified nodules, the development of which is associated with abnormal differentiation of resident fibroblasts known as valvular interstitial cells (VICs). Upon tissue injury, VICs are activated to myofibroblasts which deposit excessive collagen and stiffen the matrix. Understanding how the pathogenic phenotype of VICs is regulated by cues from the matrix may lead to new therapeutic treatments for CAS. In this thesis, I examined how matrix elasticity and TGF-&beta;1 regulate VIC phenotypes. First, I characterized the VIC population from porcine aortic valves and showed that this population is relatively homogeneous. When I cultured these primary cells on different substrates, I found that poly(ethylene glycol) hydrogels mimicked the native valve matrix better than tissue culture polystyrene plates with respect to preserving the quiescent fibroblast phenotype. At the level of signaling, I demonstrated that this is mediated through an elasticity-regulated PI3K/AKT pathway. Additionally, I showed that reduced matrix rigidity redirected activated valvular myofibroblasts into dormant fibroblasts without inducing significant apoptosis. Finally, I examined the effect of TGF-&beta;1 on VIC gene expression over time with microarray-based gene expression profiling and found that TGF-&beta;1 up-regulated cell-cell contact proteins (e.g., OB-cadherin, N-cadherin) in order to regulate valvular myofibroblast activation. Collectively, my thesis work revealed novel mechanosensing mechanisms employed by VICs to respond to matrix elasticity and explored the complex interactions among multiple extracellular cues, including matrix elasticity, TGF-&beta;1 and cell-cell adhesion, to direct the cellular fate of VICs.</p>

Les algues brunes, eucaryotes photosynthétiques marins appartenant à la lignée des Stramenopiles, présentent une trajectoire évolutive distincte de celle des plantes, des animaux et des autres eucaryotes multicellulaires. Au sein de cette lignée, elles constituent le seul groupe à avoir développé une multicellularité complexe, accompagnée de l'élaboration d'une matrice extracellulaire (MEC) entourant leurs cellules. Cette MEC, essentiellement polysaccharidique, se distingue par une biochimie unique parmi les eucaryotes, dominée par des alginates et des polysaccharides sulfatés contenant du fucose (FCSP), dont les fucanes. Malgré leur rôle essentiel dans la physiologie des algues, la compréhension des voies métaboliques de ces polysaccharides a longtemps été entravée par un manque de données génomiques. Si l'accès à la séquence génomique d'Ectocarpus en 2010 a permis d'établir les bases moléculaires, l'accès à des données génomiques plus vastes étaient nécessaires. Dans le cadre de cette thèse, j'ai adopté diverses approches pour étudier des enzymes clés impliquées dans le métabolisme de l'alginate et du fucane. Grâce à des efforts collaboratifs, combinés à l'accès à un vaste jeu de données génomiques, nous avons annoté un large panel de gènes impliqués dans ces métabolismes, dont les enzymes actives sur les carbohydrates (CAZymes). Nous avons affiné nos prédictions de gènes candidats impliqués dans la biosynthèse des fucanes, en particulier les gènes codant pour des fucosyltransférases. Nous avons retracé l'histoire évolutive du métabolisme de l'alginate au sein des Stramenopiles. J'ai étudié la multimodularité de deux familles de CAZymes potentiellement impliquées dans la dégradation de l'alginate, mettant en évidence de nouveaux extra-domaines susceptibles de jouer un rôle dans le métabolisme de l'alginate et dans les processus de signalisation. Pour compléter ces analyses in silico, j'ai exploré l'expression recombinante d'un large éventail d'enzymes clés potentiellement impliquées dans la biosynthèse des fucanes et la dégradation de l'alginate. Malgré les défis notoires associés à l'expression des protéines issues d'algues brunes, j'ai réussi à exprimer une fucosyltransférase, SlFUT1, sous forme soluble. Fort de ce succès, j'ai optimisé sa production ainsi que sa purification, et confirmé que le GDP-fucose est le donneur préférentiel de SlFUT1. J'ai également démontré son activité fucosyltransférase sur des FCSPs et des oligofucanes non sulfatés synthétisés chimiquement. La caractérisation biochimique de SlFUT1 en fait la première glycosyltransférase validée ayant une activité de synthèse du fucane. Cette recherche marque une avancée majeure dans notre compréhension des voies métaboliques des polysaccharides de la matrice extracellulaire des algues brunes, ouvrant de nouvelles perspectives pour l'étude fonctionnelle de ces métabolismes<br>Brown algae are marine eukaryotes belonging to the Stramenopiles lineage. They exhibit a distinct evolutionary trajectory from plants, animals, and other multicellular eukaryotes. They are the only group to have developed complex multicellularity within this lineage, which came with the formation of an elaborate extracellular matrix (ECM) surrounding their cells. This ECM features a unique biochemistry among eukaryotes, predominantly dominated by gel-forming alginates and fucose-containing sulfated polysaccharides (FCSPs), including fucans. Despite their critical role in algal physiology, the understanding of carbohydrate metabolic pathways in brown algae has long been hindered by a lack of genomic data. While sequencing the Ectocarpus genome in 2010 laid groundwork for molecular studies, broader genome analyses were needed. In this thesis, I adopted a range of approaches to investigate key enzymes involved in alginate and fucan metabolism. Through collaborative efforts paired with access to multiple genomic data, we successfully annotated a large range of genes involved in these metabolisms, including Carbohydrate Active Enzymes (CAZymes). We refined predictions for fucan biosynthesis, with a particular focus on enzymes encoding fucosyltransferases. We also traced the evolutionary history of alginate metabolism within Stramenopiles. I further examined the multimodular nature of two CAZyme families potentially involved in alginate degradation, uncovering novel extra-domains that may play roles in both alginate metabolism and signaling processes. To complement these in silico analyses, I explored the recombinant expression of a wide range of key enzymes involved in fucan biosynthesis and alginate degradation. Despite the notorious challenges of expressing brown algal proteins, I succeeded in expressing a putative fucosyltransferase, SlFUT1, in its soluble form. Building on this success, I optimized both its production and purification and confirmed that GDP-fucose is the preferred sugar donor for SlFUT1. I also demonstrated its fucosyltransferase activity on both FCSPs and unsulfated oligofucans which were chemically synthesized. The biochemical characterization of SlFUT1 establishes it as the first validated glycosyltransferase with a fucan-synthase activity. This research advances our understanding on the metabolic pathways of extracellular matrix polysaccharides in brown algae, paving the way for further functional exploration of these metabolisms

Giant settlements worldwide which incorporated large amounts of open space, such as Cahokia, Great Zimbabwe and the European oppida, have long seemed anomalous to scholarship because they did not match the characteristics of conventional urbanism. It is now apparent that these settlements appeared relatively frequently in diverse conditions across the world over the past 7000 years and they may constitute a form of human settlement behaviour that has not yet been consistently articulated. In Limits of Settlement Growth, Roland Fletcher (1995) identified that such settlements sit outside usual categories for specifying conditions of settlement operation in the past. He predicted that these settlements would have dropped to a low-density internal pattern as they expanded to large size. In his parlance they operated below a Threshold Limit on the Interaction-Communication Matrix and would not therefore be constrained in their areal expansion. The aim of this research was to test Fletcher’s prediction for these sites in the I-C Model and explore the findings to delineate boundary conditions of their operation. A dataset of sites 100 ha or larger was compiled and compared across variables such as area, population mobility, durations and demise. Using a recently updated version of the I-C Matrix, these settlements were found to have operated beneath the Mobile Interaction-Limit as well as at lower densities. The mobility component of their operation varied substantially and there may be more than one type of human settlement behaviour indicated. Structurally, they were giant variants of preceding and contemporaneous settlement forms emergent under conditions of regional or sub-regional population increase rather than transformations. While individually they could have short durations, they could appear multiple times in a culture region and were resilient to all but regional systemic change.

Cette thèse a pour objectif de développer une méthode de mesure de champs par corrélation d’images numériques (CIN) à haute température couplée à des mesures thermiques sur une éprouvette technologique en CMC sollicitée dans des conditions thermiques représentatives d’un environnement moteur et de mettre en place une méthodologie d’identification des propriétés thermiques et thermomécaniques du matériau, en quantifiant à chaque étape de la chaîne les incertitudes associées aux quantités d’intérêt et en les réduisant. Il a pour cela été nécessaire de traiter les défis inhérents à la CIN à chaud, que ce soit au niveau de l’acquisition des images (saturation, perte du contraste) ou de la mesure (artefacts dus à l’effet mirage, aussi appelé "brume de chaleur").Ces travaux ont ainsi donné lieu au développement d’un protocole d’étalonnage d’un banc multi-instrumenté par l’utilisation soit d’une mire in-situ, soit par auto-étalonnage en utilisant l’éprouvette elle-même et son environnement. Les mesures de déplacements 3D surfaciques (approches de stéréocorrélation globales) et les mesures thermiques ont permis de mettre en évidence ce phénomène de brume de chaleur. Des stratégies de régularisation spatiotemporelles des déplacements mesurés ont été proposées et ont permis d’obtenir des résultats satisfaisants (réduction significative des incertitudes de mesure). De même, des approches par réduction de modèles (POD) ont permis de traiter les données thermiques et de quantifier les incertitudes associées aux phénomènes convectifs. Enfin, un algorithme de recalage de modèle éléments finis pondéré sur les données de températures et de déplacements a été implémenté en vue d’identifier un ensemble de propriétés thermiques et thermomécaniques, en tenant compte de la sensibilité de chaque paramètre par rapport aux incertitudes de mesures<br>The aim of this thesis is firstly to develop procedures of full-field measurements with Digital Image Correlation (DIC), coupled to thermal measurements, suitable for high-temperature experiments on CMC specimens under thermal conditions representative of an engine environment. Secondly, a methodology is proposed for identifying the thermal and thermomechanical properties of the material, quantifying at each stage of the chain the uncertainties associated with the quantities of interest and strategies to reduce them. It was necessary to deal with the challenges due to high temperatures, especially for DIC, either in terms of acquisition (saturation, loss of contrast) or measurement (artefacts due to the mirage effect, also called "heat haze effect").This work has led to the development of a calibration protocol for a multi-instrumented bench using either an in-situ calibration target or by self-calibration using the specimen itself and its environment. 3D surface displacement measurements (with global stereocorrelation approaches) and thermal measurements have made it possible to highlight the heat haze effect phenomenon. Spatiotemporal regularisation strategies of the measured displacements were proposed and allowed satisfactory results to be obtained (significant reduction of measurement uncertainties). Similarly, model reduction approaches (POD) have been used to process thermal data and quantify the uncertainties associated with convective phenomena. Finally, a weighted Finite-Element Model Updating (FEMU) algorithm on both temperature and displacement data was implemented in order to identify a set of thermal and thermomechanical properties, taking into account the sensitivity of each parameter with regard to measurement uncertainties

The present thesis reviews my scientific works on disordered systems from 1995 until today. They can be roughly categorized into three main classes: (1) non-interacting disordered systems, (2) the two-interacting particle problem, and (3) the interplay of disorder and many-particle interaction. A (4)th chapter is concerned with the implementation of the numerical algorithms. The structure of the thesis reflects this division. The reprints have been added at the end of these main divisions according to their context. For the convenience of the reader, I have ordered them in each chapter alphabetically according to the names of the authors. Furthermore, in each citation of my work, the starting page number in the thesis is given, e.g, Ref.\ \cite{EPR97} refers to a paper of Eckle, Punnoose and myself and can be found on page \pageref{EPR97}. Citations which do not refer to my work are numbered and are ordered in the bibliography according to the names of the authors.

Process models have been used to predict the flow and stress development during manufacturing of composite structures for over 20 years. To date, these models have been treated separately, and the required material property models (viscosity and modulus) have also been treated separately. The latest breakthrough in process modeling of composite structures is an integrated stress-flow model. However, a consistent viscosity and viscoelastic material model required by the integrated stress-flow model has not been developed. Presented in this thesis is a consistent material model that predicts the viscoelastic liquid and viscoelastic solid behavior of a commercially available thermoset polymer, namely MTM45-1 epoxy. The goal here is to show that a single material model can predict the viscosity and viscoelastic modulus for all temperature, degree of cure and time scales encountered in composite manufacturing. The model was generated by fitting a generalized Maxwell model to test results from both dynamic mechanical analysis and rheological tests. Both resin and prepreg samples were examined. Thermo-rheological complex behavior was captured by applying linear temperature dependence to the un-relaxed modulus. The effect of cure was accounted for by applying a degree of cure dependent shift function. The relaxed modulus was predicted using the cross-link concentration and the theory of rubber. Excellent agreement was found when comparing predictions from the model to experimental data ranging from temperatures of -50°C to 245°C, degree of cure of 0.01 to 1.0 and frequencies of 0.01Hz to 10Hz.

Abstract Introduction The demand for organ transplantation has rapidly increased during the past decades, thus requiring the development of a new interdisciplinary field, aimed at supplying this demand. Branching from regenerative medicine, the combination of elements usually applied separately for same or other purposes, was named Tissue Engineering (TE). Precisely, the components combined in TE constitute the so-called TE triad, are i) cells (derived from cell therapy), ii) scaffolding material (derived from material science) and iii) molecular signals (mainly derived from molecular biology and drug research). Among the organs and tissue that experience damages such that requiring implantation/transplantation, skeletal muscle represents no exception. As result of this, several approaches developed to efficiently repair the most common defects requiring surgery: Volumetric muscle loss (VML), abdominal wall defects (AWD) and defects of the diaphragm, namely traumatic diaphragmatic hernias or congenital diaphragmatic hernias (CDHs). While the first necessitates a substitute large in volume, the other two need a thinner but, according to defect size, wider surface. Thus far, between the materials chosen for repair of such defects, mostly synthetic materials were used, but decellularized tissue are rapidly covering the gap, likely to surpass them in the next future. Reason of this potential is the close resemblance to the organ or tissue of origin, while possessing features able to endpoint to constructive and functional remodeling. While decellularization-derivates of several organs and tissue were already attained, some were even tested in clinic and further moved to large-scale production (i.e. Surgisis®). However, only recently the proposal of developing a decellularized scaffold from diaphragm muscle for autologous repair purposes was reprised, and in general, was never considered in mouse. Methods Diaphragms (mouse and rabbit) were decellularized by a cyclical exposure to the sequence deionised water, sodium deoxycholate, DNase. Scaffolds were characterized in order to evaluate decellularization protocol efficiency, by the means of cell removal (mouse and rabbit), maintenance of microarchitecture (mouse and rabbit), ECM components (mouse), mechanical properties (mouse). Next, the interaction host-scaffold was tested in vivo, both in healthy and atrophic mice. Subsequent, a further characterization by disclosing the angiogenetic properties, both potential, taking advantage of the well-known CAM assay and subcutaneous transplant, as well as revealing angiogenesis-related protein content whithin the scaffold, and actual, by performing an orthotropic transplant in comparison with a synthetic material. To draw nearer the clinical validation, the characterized decellularized diaphragm (DD) were thus transplanted in the first surgically created model of CDH, again compared to the most commonly used material for the repair of this defect, polytetrafluoroethylene (PTFE), with the introduction recently of an allotransplant control. Lastly, as completion of a TE approach, it was developed a method to combine the cellular component, which was then applied to both mouse and rabbit derived scaffold. Results The adaption of a previously published detergent-enzymatic protocol aiming at obtaining a decellularized scaffold from diaphragm, resulted successful in with both mouse and rabbit sources, as cell removal while preserving microarchitecture and ECM components was attained. Moreover, it was confirmed that mechanical properties and growth factors were preserved in the mouse-derived scaffold. Upon implantation of the latter, angiogenesis and myogenic activation, while modulating the immune response were observed. In the healthy environment, the effect was transient, whereas implantation in a mouse model of atrophy led to long-term beneficial effects. The testing of mouse scaffold in a surgical model of CDH proved the previously seen results, as, compared to PTFE, yielded better outcomes, such as no recurrence, amelioration of diaphragm excursion and, differently from before, to a sustained myogenic response through time points analysed. The method used to re-populate both mouse and acellular muscles , after being developed, thus far resulted in a successful cell delivery, while the scaffold supported cell proliferation, survival and differentiation. Conclusions Muscle scaffold from mice and rabbit can be successfully obtained by decellularizing the diaphragm muscle via detergent enzymatic protocol. These scaffolds were proven to have several attractive properties both in vitro and in vivo, derived from their close resemblance to the tissue of origin. Precisely, it was demonstrated that while in vitro can sustain cell survival, proliferation and differentiation, in vivo positively interact with the recipient environment, driving a constructive response. Hence, scaling up of this type of construct is likely to be happening in the next future.<br>Riassunto Introduzione Negli ultimi decenni, la lista di attesa dei pazienti che necessitano trapianto di organi o un intervento chirurgico mirato alla sostituzione di grandi porzioni di tessuto, è andata aumentando. Al contrario però, la disponibilità di donatori, sia vivi che cadaverici, non ha seguito lo stesso andamento. Come conseguenza, nonostante i tentativi di supplire a queste necessità, molti ancora muoiono in attesa di essere curati. La medicina rigenerativa, ovvero la combinazione di diversi elementi dell’ingegneria tissutale, nasce proprio per rispondere alla ancora pressante domanda di organi e tessuti trapiantabili. Precisamente, un approccio di medicina rigenerativa si avvale di tre componenti principali: i) le cellule (dalla terapia cellulare), ii) il supporto o impalcatura per favorirne la crescita (derivato dalla scienza dei materiali, ad esempio prostetici, comunemente chiamato scaffold o mesh) e iii) segnali molecolari (principalmente dalla biologia molecolare e della farmaceutica). Tra gli organi e tessuti che sperimentano danni tali da richiedere l'impianto/trapianto, vi è anche il muscolo scheletrico. Di conseguenza, anche per questo sono stati sviluppati diversi approcci mirati a riparare efficacemente i più comuni difetti sia pediatrici sia adulti che richiedono un intervento chirurgico: A) la perdita muscolare di un grande volume di muscolo (causato da incidenti o dalla necessità di rimuovere tumori muscolari), B) difetti della parete addominale e C) difetti del diaframma, soprattutto congeniti. A seconda del tipo di difetto, il sostituto da sviluppare dovrà avere peculiari proprietà, oltre a quelle basilari necessarie per essere considerato un approccio di ingegneria tissutale completo. Infatti, mentre A necessita di un sostituto in grado di ripristinare un grande volume, B e C richiedono un costrutto sottile ma, a seconda delle dimensioni del difetto, di una più o meno ampia superficie. Finora, per la riparazione di tali difetti per lo più sono stati utilizzati più frequentemente scaffold/mesh sintetici, ma quelli di derivazione naturale, ed in particolare ottenuti dalla decellularizzazione di un tessuto nella sua completezza, stanno rapidamente recuperando strada, ed è probabile che diverranno il metodo di elezione nel prossimo futuro. La ragione di questo potenziale è la stretta somiglianza con l'organo o il tessuto di origine, pur possedendo caratteristiche in grado di culminare in un rimodellamento costruttivo e funzionale. Inoltre, mentre la decellularizzazione è stata ottenuta anche in organi più complessi, i prodotti derivati da tessuti più semplici sono stati anche già testati in clinica e per alcuni si è persino giunti alla commercializzazione su larga scala (ad esempio il Surgisis®). In questo lavoro si è partiti dal considerare la patologia dell’ernia diaframmatica, ad oggi riparata chirurgicamente tramite chiusura primaria con tessuto autologo, se il difetto lo permette, oppure ancora con mesh prostetici di natura sintetica. Tuttavia, gli svantaggi di questi tipi di materiali in termini di rigidità ed inerzia, rendono conto dell’attuale ricerca di una valida alternativa. Si è proposto quindi di derivare un costrutto muscolare partendo dalla decellularizzazione del muscolo diaframmatico utilizzabile come toppa tissutale per riparare il danno dell’ernia sia senza cellule sia con nuove cellule donatrici. Il modello studiato è stato, per la prima volta, quello murino. Metodi Diaframmi (topo e coniglio) sono stati decellularizati tramite esposizione ciclica alla sequenza: acqua deionizzata, sodio desossicolato, DNasi. Gli scaffold ottenuti sono stati caratterizzati per valutare l'efficienza del protocollo in termini di rimozione delle cellule (topo e coniglio), preservazione della micro architettura (topo e coniglio), componenti della matrice extracellulare (topo), proprietà meccaniche (topo). Successivamente, l'interazione ospite-scaffold (in topo) è stata testata in vivo, sia in topi sani che atrofici. Un ulteriore caratterizzazione è stata attuata approfondendo le proprietà angiogeniche, sia potenziali, sfruttando il noto test CAM, il trapianto sottocutaneo e rivelando il contenuto di proteine coinvolte nella angiogenesi ancora presenti dopo la decellularizzazione, sia effettive, eseguendo un impianto ortotopico, utilizzando come confronto un materiale sintetico. Per avvicinare la validazione alla pratica clinica, il diaframma decellularizato di topo precedentemente caratterizzato è stato quindi utilizzato per il riparo nel primo modello di ernia diaframmatica chirurgicamente indotta in topo. Il paragone è stato fatto anche in questo caso con il materiale più comunemente usato per la riparazione di questo difetto in clinica. Inoltre, più recentemente è stato introdotto un controllo rappresentato dall’allotrapianto. Infine, a completamento di un approccio di ingegneria tissuatle, è stato sviluppato un metodo per combinare la componente cellulare, poi applicato a entrambi gli scaffold derivati da topo e coniglio. Risultati L'adattamento di un protocollo detergente enzimatico precedentemente pubblicato, al fine di ottenere uno scaffold decellularizato da diaframma, si è rivelato efficiente sia utilizzando il topo che il coniglio come fonti, ottenendo così da entrami la rimozione delle cellule preservando la somiglianza della struttura con il tessuto di origine. Inoltre, è stato confermato, solo in topo, che le proprietà meccaniche e i fattori di crescita sono stati conservati dopo la decellularizzazione. L'impianto dello scaffold di topo, ha provocato sia angiogenesi che l'attivazione dei precursori muscolari, modulando al contempo la risposta immunitaria. Nell'ambiente sano, l'effetto è stato transitorio, mentre l'impianto in un modello murino di atrofia ha portato ad effetti benefici a lungo termine. La sperimentazione nel modello chirurgico di ernia ha confermato i risultati precedentemente osservati ed in aggiunta, rispetto al materiale sintetico, ha mostrato migliori risultati, come l'assenza di erniazione recidiva, il miglioramento dell’escursione diaframmatica e, diversamente da prima, una più sostenuta risposta miogenica nel tempo. Il metodo utilizzato per ripopolare entrambi i diaframmi acellulari di topo e muscoli, dopo essere stato sviluppato, ha finora portato a risulstati positivi, mentre lo scaffold ha dimostrato di poter supportare la proliferazione, la sopravvivenza e la differenziazione delle cellule. Conclusioni Scaffold derivati dal muscolo di diaframma possono essere conseguiti con successo sia da topi che conigli, utilizzando un protocollo di tipo detergente-enzimatico. Tali, hanno dimostrato di possedere diverse proprietà interessanti sia in vitro che in vivo, derivate dalla loro stretta somiglianza con il tessuto di origine. Precisamente, è stato dimostrato che, mentre in vitro possono sostenere la sopravvivenza, la proliferazione e la differenziazione cellulare, in vivo sono in grado di interagire positivamente con l'ambiente ricevente, guidando una risposta costruttiva. Quindi, è molto probabile che ulteriori passi vengano fatti nel prossimo futuro, fino a raggiungere infine l’applicazione in clinica.

Master of Science<br>Department of Civil Engineering<br>Hayder A. Rasheed<br>Throughout history natural fiber was used as one of the main building materials all over the world. Because the use of such materials has decreased in the last century, not much research has been conducted to investigate their performance as a reinforcing material in cement and concrete. In order to investigate one of the most common natural fibers, wheat fibers, as a reinforcing material, 156 mortar specimens and 99 concrete specimens were tested. The specimens were tested in either uniaxial compression or flexure. The uniaxial compression test included 2 in (50.8 mm) mortar cubes and 4x8 in (101.6 x 203.2 mm) concrete cylinders. As for the flexure test, they were either 40x40x160 mm cementitious matrix prisms or 6x6x21 in (152.4x152.4x533.4 mm) concrete prisms. Several wheat fibers percentages were studied and compared with polypropylene fiber as a benchmarking alternative. The average increase in the uniaxial compression strength for cementitious matrix cubes reinforced with 0.5% long wheat fiber exceeded that of their counterparts reinforced with polypropylene fiber by 15%. Whereas for concrete cylinders reinforced with 0.75% long wheat fiber, their strength exceeded that of their counterparts reinforced with polypropylene fiber by 5% and that of the control by 7%. The flexural strength of cementitious matrix prisms reinforced with 0.75% long wheat fiber exceeded that of their counterparts reinforced with polypropylene fiber by 27%. Meanwhile, concrete prisms reinforced with both long wheat fiber and polypropylene fiber showed deterioration in strength of up to 17%. Finally, ABAQUS models were developed for concrete cylinders and prisms to simulate the effect of inclusion of the wheat fibers.

Membranes are a promising, effective and energy efficient separation strategy for effluent gases in the Reverse Water Gas Shift (RWGS) reaction to increase the overall conversion of CO2 to CO. This process involves a separation and recycling process to reuse the unreacted CO2 from the RWGS reactor. The carbon monoxide produced from this reaction, alongside hydrogen (composing syngas), can be used in the Fischer-Tropsch process to create synthetic fuel, turning stationary CO2 emissions into a useable resource. A literature review was performed to select suitable polymers with high CO2 permeability and selectivities of CO2 over CO and H2. PDMS (polydimethylsiloxane) was selected and commercial and in-house PDMS membranes were tested. The highest CO2 permeability observed was 5,883 Barrers, including a CO2/H2 selectivity of 21 and a CO2/CO selectivity of 9, with ternary gas feeds. HY zeolite, silica gel and activated carbon were selected from previous research for their CO2 separation capabilities, to be investigated in PDMS mixed matrix membranes in 4 wt % loadings. Activated carbon in PDMS proved to be the best performing mixed matrix membrane with a CO2 permeability of 2,447 Barrers and comparable selectivities for CO2/H2 and CO2/CO of 14 and 9, respectively. It was believed that swelling, compaction and the homogeneity of the selective layer were responsible for trends in permeability with respect to driving force. The HY and silica gel mixed matrix PDMS membranes were believed to experience constraints in performance due to particle and polymer interfaces within the membrane matrix.

L’intérêt croissant de l’industrie pour les matériaux composites thermoplastiques est motivé par leurs propriétés de thermoformabilité, de recyclabilité ainsi que leurs capacités de cadences de production élevées. Le développement de matériaux pré-imprégnés thermoplastiques, apparus dès les années 1980, s’est imposé comme un moyen efficace de contourner les fortes viscosités des polymères utilisés en réduisant la distance d’écoulement des polymères à l’état « fondu ». Cette étude s’est plus particulièrement intéressée au développement de composites à base de tissus de verre et de carbone pré-imprégnés par un latex acrylique, le TPREG I. En outre, les propriétés mécaniques élevées des matrices acryliques, alliées à un coût relativement faible, en font un matériau intéressant, de nature à permettre un saut technologique dans la conception et la fabrication de composites structuraux à matrice organique. Notre étude s’est concentrée sur la mesure de l’adhésion à l’interface fibre/matrice acrylique car cette région est au cœur du transfert de charge de la matrice vers les fibres et conditionne donc les propriétés mécaniques du composite. Nous avons choisi d’évaluer l’adhésion interfaciale en combinant des analyses de mouilllage avec des tests mécaniques aux échelles microscopique et macroscopique. Le test micromécanique de la microgoutte permet de mettre en évidence le rôle central de l’ensimage des fibres sur la contrainte de cisaillement interfaciale. L’adhésion thermodynamique, déterminé par des mesures d’énergie de surface, est en accord avec la contrainte de cisaillement et souligne l’influence de la polarité de l’ensimage. A l’échelle macroscopique, les essais de traction hors-axe sur composites unidirectionnels permettant de solliciter l’interface en cisaillement quasi-plan ont mis en exergue une corrélation entre les échelles micro et macro. L’étude a également permis de dégager une forte augmentation de l’adhésion grâce à une modification de la matrice acrylique, ainsi qu’une dégradation des propriétés interfaciales à l’échelle micro par vieillissement hydrolytique. Cette étude constitue une première base de données concernant les propriétés interfaciales de composites thermoplastiques acryliques et démontre l’importance d’une étude multi-échelles dans la conception de nouveaux composites<br>The present study was initiated by the development of a new processing route, i.e. latex-dip impregnation, for thermoplastic (TP) acrylic semi-finished materials. The composites resulting from thermocompression of TPREG I plies were studied by focusing of interfacial adhesion. Indeed the fiber/matrix interface governs the stress transfer from matrix to fibers. Thus, a multi-scale analysis of acrylic matrix/fiber interfaces was conducted by considering microcomposites, as models for fiber-based composites, and unidirectional (UD)macro-composites. The study displayed various types of sized glass and carbon fibers. On one hand, the correlation between thermodynamic adhesion and practical adhesion, resulting from micromechanical testing, is discussed by highlighting the role of the physico-chemistry of the created interphase. Wetting and thermodynamical adhesion are driven by the polarity of the film former of the sizing. On the other hand, in-plane shear modulus values from off-axis tensile test results on UD composites are consistent with the quantitative analyses of the interfacial shear strength obtained from microcomposites. More specifically, both tests have enabled a differentiation of interface properties based on the fiber sizing nature for glass and carbon fiber-reinforced (micro-)composites. The study of overall mechanical and interface properties of glass and carbon fiber/acrylic composites revealed the need for tailoring interfacial adhesion. Modifications of the matrix led to successful increases of interfacial adhesion in glass fiber/acrylic composites. An additional hygrothermal ageing study evidenced a significant loss of interfacial shear strength at micro-scale which was not observed for UD composites. The results of this study are a first step towards a database of relevant interface properties of structural TP composites. Finally, the analyses of interfaces/phases at different scales demonstrate the importance of a multi-scale approach to tailor the final properties of composite parts

Les travaux présentés dans cette thèse relèvent de l'étude des méthodes de classification linéaires, c'est à dire l'étude de méthodes ayant pour but la catégorisation de données en différents groupes à partir d'un jeu d'exemples, préalablement étiquetés, disponible en amont et appelés ensemble d'apprentissage. En pratique, l'acquisition d'un tel ensemble d'apprentissage peut être difficile et/ou couteux, la catégorisation d'un exemple étant de fait plus ardu que l'obtention de dudit exemple. Cette disparité entre la disponibilité des données et notre capacité à constituer un ensemble d'apprentissage étiqueté a été un des problèmes centraux de l'apprentissage automatique et ce manuscrit s’intéresse à deux solutions usuellement considérées pour contourner ce problème : l'apprentissage en présence de données bruitées et l'apprentissage actif<br>The works presented in this thesis fall within the general framework of linear classification, that is the problem of categorizing data into two or more classes based on on a training set of labelled data. In practice though acquiring labeled examples might prove challenging and/or costly as data are inherently easier to obtain than to label. Dealing with label scarceness have been a motivational goal in the machine learning literature and this work discuss two settings related to this problem: learning in the presence of noise and active learning

La manière dont l'information sensorielle est codée et traitée par les circuits neuronaux est une question centrale en neurosciences computationnelles. Dans de nombreuses régions du cerveau, on constate que l'activité des neurones dépend fortement de certains corrélats sensoriels continus ; on peut citer comme exemples les cellules simples de la zone V1 du cortex visuel codant pour l'orientation d'une barre présentée à la rétine, et les cellules de direction de la tête dans le subiculum ou les cellules de lieu dans l'hippocampe, dont les activités dépendent, respectivement, de l'orientation de la tête et de la position d'un animal dans l'espace physique. Au cours des dernières décennies, les réseaux neuronaux à attracteur continu ont été introduits comme un modèle abstrait pour la représentation de quelques variables continues dans une grande population de neurones bruités. Grâce à un ensemble approprié d'interactions par paires entre les neurones, la dynamique du réseau neuronal est contrainte de s'étendre sur une variété de faible dimension dans l'espace de haute dimension des configurations d'activités, et code ainsi quelques coordonnées continues sur la variété, correspondant à des informations spatiales ou sensorielles. Alors que le modèle original était basé sur la construction d'une variété continue unique dans un espace à haute dimension, on s'est vite rendu compte que le même réseau neuronal pouvait coder pour de nombreux attracteurs distincts, correspondant à différents environnements spatiaux ou situations contextuelles. Une solution approximative à ce problème plus difficile a été proposée il y a vingt ans, et reposait sur une prescription ad hoc pour les interactions par paires entre les neurones, résumant les différentes contributions correspondant à chaque attracteur pris indépendamment des autres. Cette solution souffre cependant de deux problèmes majeurs : l'interférence entre les cartes limitent fortement la capacité de stockage, et la résolution spatiale au sein d'une carte n'est pas contrôlée. Dans le présent manuscrit, nous abordons ces deux questions. Nous montrons comment parvenir à un stockage optimal des attracteurs continus et étudions le compromis optimal entre capacité et résolution spatiale, c'est-à-dire comment l'exigence d'une résolution spatiale plus élevée affecte le nombre maximal d'attracteurs pouvant être stockés, prouvant que les réseaux neuronaux récurrents sont des dispositifs de mémoire très efficaces capables de stocker de nombreux attracteurs continus à haute résolution. Afin de résoudre ces problèmes, nous avons utilisé une combinaison de techniques issues de la physique statistique des systèmes désordonnés et de la théorie des matrices aléatoires. D'une part, nous avons étendu la théorie de l'apprentissage de Gardner au cas des modèles présentant de fortes corrélations spatiales. D'autre part, nous avons introduit et étudié les propriétés spectrales d'un nouvel ensemble de matrices aléatoires, c'est-à-dire la superposition additive d'un grand nombre de matrices aléatoires euclidiennes indépendantes dans le régime de haute densité. En outre, cette approche définit un cadre concret pour répondre à de nombreuses questions, en lien étroit avec les expériences en cours, liées notamment à la discussion de l'hypothèse du remapping aléatoire et au codage de l'information spatiale et au développement des circuits cérébraux chez les jeunes animaux. Enfin, nous discutons d'un mécanisme possible pour l'apprentissage des attracteurs continus à partir d'images réelles<br>How sensory information is encoded and processed by neuronal circuits is a central question in computational neuroscience. In many brain areas, the activity of neurons is found to depend strongly on some continuous sensory correlate; examples include simple cells in the V1 area of the visual cortex coding for the orientation of a bar presented to the retina, and head direction cells in the subiculum or place cells in the hippocampus, whose activities depend, respectively, on the orientation of the head and the position of an animal in the physical space. Over the past decades, continuous attractor neural networks were introduced as an abstract model for the representation of a few continuous variables in a large population of noisy neurons. Through an appropriate set of pairwise interactions between the neurons, the dynamics of the neural network is constrained to span a low-dimensional manifold in the high-dimensional space of activity configurations, and thus codes for a few continuous coordinates on the manifold, corresponding to spatial or sensory information. While the original model was based on how to build a single continuous manifold in an high-dimensional space, it was soon realized that the same neural network should code for many distinct attractors, {em i.e.}, corresponding to different spatial environments or contextual situations. An approximate solution to this harder problem was proposed twenty years ago, and relied on an ad hoc prescription for the pairwise interactions between neurons, summing up the different contributions corresponding to each single attractor taken independently of the others. This solution, however, suffers from two major issues: the interference between maps strongly limit the storage capacity, and the spatial resolution within a map is not controlled. In the present manuscript, we address these two issues. We show how to achieve optimal storage of continuous attractors and study the optimal trade-off between capacity and spatial resolution, that is, how the requirement of higher spatial resolution affects the maximal number of attractors that can be stored, proving that recurrent neural networks are very efficient memory devices capable of storing many continuous attractors at high resolution. In order to tackle these problems we used a combination of techniques from statistical physics of disordered systems and random matrix theory. On the one hand we extended Gardner's theory of learning to the case of patterns with strong spatial correlations. On the other hand we introduced and studied the spectral properties of a new ensemble of random matrices, {em i.e.}, the additive superimposition of an extensive number of independent Euclidean random matrices in the high-density regime. In addition, this approach defines a concrete framework to address many questions, in close connection with ongoing experiments, related in particular to the discussion of the random remapping hypothesis and to the coding of spatial information and the development of brain circuits in young animals. Finally, we discuss a possible mechanism for the learning of continuous attractors from real images

The intensification of agriculture for increased food production is leading to new challenges for biodiversity conservation, particularly managing complex changing landscapes for mutually beneficial outcomes for agriculture and the environment. My thesis aimed to understand the diversity and distribution of beetles (Coleoptera), and the mechanisms shaping beetle assemblages across a dynamic and fragmented agricultural landscape. I used a landscape-scale study in south-eastern Australia to examine beetle assemblages in remnant woodland patches and four types of adjoining farmlands: crop, fallow, plantings, and fine woody debris applied over harvested crop. My thesis comprised four chapters written as journal articles. In Paper I, I examined seasonal differences in beetle assemblages between the woodland interior and four adjoining farmland uses. I found that overall species richness was significantly lower in woodlands than farmlands, although both habitats supported significantly different assemblages. Abundance responses were taxon-specific, and influenced by interactions between land-use and season. These results suggest the importance of maintaining farmland heterogeneity with a mix of low-intensity land-uses, with further agricultural intensification a likely threat to beetle diversity in the region. In Paper II, I examined temporal patterns of edge responses and movement of beetle assemblages between woodlands and the four farmland uses. The use of directional pitfall traps allowed inference of cross-habitat movement. Farmland use and season interactively affected beetle abundance across farmland–woodland edges. Applying woody debris was a novel way of reducing seasonal fluctuations in edge responses and increasing permeability for cross-habitat movement. Edges likely provided resources for beetles in adjoining habitats, but seasonal movement of predators into edges might negatively affect prey assemblages. In Paper III, I quantified relationships between ground-layer structure, plant species richness and plant composition, and the diversity and composition of beetles from different habitats or seasons. Plant composition better predicted beetle composition than vegetation structure. Plant richness and vegetation structure both significantly affected beetle abundance and composition. The influence of these vegetation attributes often varied depending on habitat and season for all trophic groups. These dynamic plant–beetle relationships suggest a need for targeted ways of managing vegetation to improve beetle diversity in different parts of the landscape. In Paper IV, I disentangled the effects of farmland use, edge effects and vegetation structure on the morphological traits of Carabidae species. Carabid body size increased across a distance from edges between woodlands and farmlands, and there were strong mediating effects of farmland use on this association. Vegetation structure was associated with traits relating to body size, flying ability and body shape, and helped explain some of the effects of farmland use and edge effects on body size. These results provide evidence of vegetation- and land-use-mediated filtering of traits as an important factor shaping carabid assemblages in human-modified landscapes. My results indicate that farmlands can provide important habitat for many beetle species. However, spatio-temporal changes in farmland habitat strongly influence beetle assemblages across the landscape. Conservation strategies, therefore, need to take a whole-of-landscape approach, and exploit heterogeneity of mixed-farmlands over space and time to maximise outcomes for biodiversity.

Dans ce manuscrit de thèse, nous introduisons les avancées récentes sur la décomposition en matrices (et tenseurs) de rang faible et parcimonieuse ainsi que les contributions pour faire face aux principaux problèmes dans ce domaine. Nous présentons d’abord un aperçu des méthodes matricielles et tensorielles les plus récentes ainsi que ses applications sur la modélisation d’arrière-plan et la segmentation du premier plan. Ensuite, nous abordons le problème de l’initialisation du modèle de fond comme un processus de reconstruction à partir de données manquantes ou corrompues. Une nouvelle méthodologie est présentée montrant un potentiel intéressant pour l’initialisation de la modélisation du fond dans le cadre de VSI. Par la suite, nous proposons une version « double contrainte » de l’ACP robuste pour améliorer la détection de premier plan en milieu marin dans des applications de vidéo-surveillance automatisées. Nous avons aussi développé deux algorithmes incrémentaux basés sur tenseurs afin d’effectuer une séparation entre le fond et le premier plan à partir de données multidimensionnelles. Ces deux travaux abordent le problème de la décomposition de rang faible et parcimonieuse sur des tenseurs. A la fin, nous présentons un travail particulier réalisé en conjonction avec le Centre de Vision Informatique (CVC) de l’Université Autonome de Barcelone (UAB)<br>This thesis introduces the recent advances on decomposition into low-rank plus sparse matrices and tensors, as well as the main contributions to face the principal issues in moving object detection. First, we present an overview of the state-of-the-art methods for low-rank and sparse decomposition, as well as their application to background modeling and foreground segmentation tasks. Next, we address the problem of background model initialization as a reconstruction process from missing/corrupted data. A novel methodology is presented showing an attractive potential for background modeling initialization in video surveillance. Subsequently, we propose a double-constrained version of robust principal component analysis to improve the foreground detection in maritime environments for automated video-surveillance applications. The algorithm makes use of double constraints extracted from spatial saliency maps to enhance object foreground detection in dynamic scenes. We also developed two incremental tensor-based algorithms in order to perform background/foreground separation from multidimensional streaming data. These works address the problem of low-rank and sparse decomposition on tensors. Finally, we present a particular work realized in conjunction with the Computer Vision Center (CVC) at Autonomous University of Barcelona (UAB)

Structural colours arise from the interaction of visible light with nano-structured materials. The occurrence of such structures in nature has been known for over a century, but it is only in the last few decades that the study of natural photonic structures has fully matured due to the advances in imagining techniques and computational modelling. Even though a plethora of different colour-producing architectures in a variety of species has been investigated, a few significant questions are still open: how do these structures develop in living organisms? Does disorder play a functional role in biological photonics? If so, is it possible to say that the optical response of natural disordered photonics has been optimised under evolutionary pressure? And, finally, can we exploit the well-adapted photonic design principles that we observe in Nature to fabricate functional materials with optimised scattering response? In my thesis I try to answer the questions above: I microscopically investigate $\textit{in vivo}$ the growth of a cuticular multilayer, one of the most common colour-producing strategies in nature, in the green beetles $\textit{Gastrophysa viridula}$ showing how the interplay between different materials varies during the various life stages of the beetles; I further investigate two types of disordered photonic structures and their biological role, the random array of spherical air inclusions in the eggshells of the honeyguide $\textit{Prodotiscus regulus}$, a species under unique evolutionary pressure to produce blue eggs, and the anisotropic chitinous network of fibres in the white beetle $\textit{Cyphochilus}$, the whitest low-refractive index material; finally, inspired by these natural designs, I fabricate and study light transport in biocompatible highly-scattering materials.

This research contributes to a very active body of literature on the application of Automated Data Collection Systems (ADCS) and openly shared data to public transportation planning. It also addresses the interaction between transit demand and land use patterns, a key component of generating time-varying origin-destination (O-D) matrices at a route level. An origin-destination (O-D) matrix describes the travel demand between two different locations and is indispensable information for most transportation applications, from strategic planning to traffic control and management. A transit passenger's O-D pair at the route level simply indicates the origin and destination stop along the considered route. Observing existing land use types (e.g., residential, commercial, institutional) within the catchment area of each stop can help in identifying existing transit demand at any given time or over time. The proposed research addresses incorporation of an alighting probability matrix (APM) - tabulating the probabilities that a passenger alights at stops downstream of the boarding at a specified stop - into a time-varying O-D estimation process, based on the passenger's trip purpose or activity locations represented by the interactions between transit demand and land use patterns. In order to examine these interactions, this research also uses a much larger dataset that has been automatically collected from various electronic technologies: Automated Fare Collection (AFC) systems and Automated Passenger Counter (APC) systems, in conjunction with other readily available data such as Google's General Transit Feed Specification (GTFS) and parcel-level land use data. The large and highly detailed datasets have the capability of rectifying limitations of manual data collection (e.g., on-board survey) as well as enhancing any existing decision-making tools. This research proposes use of Google's GTFS for a bus stop aggregation model (SAM) based on distance between individual stops, textual similarity, and common service areas. By measuring land use types within a specified service area based on SAM, this research helps in advancing our understanding of transit demand in the vicinity of bus stops. In addition, a systematic matching technique for aggregating stops (SAM) allows us to analyze the symmetry of boarding and alightings, which can observe a considerable passenger flow between specific time periods and symmetry by time period pairs (e.g., between AM and PM peaks) on an individual day. This research explores the potential generation of a time-varying O-D matrix from APC data, in conjunction with integrated land use and transportation models. This research aims at incorporating all valuable information - the time-varying alighting probability matrix (TAPM) that represents on-board passengers' trip purpose - into the O-D estimation process. A practical application is based on APC data on a specific transit route in the Minneapolis - St. Paul metropolitan area. This research can also provide other practical implications. It can help transit agencies and policy makers to develop decision-making tools to support transit planning, using improved databases with transit-related ADCS and parcel-level land use data. As a result, this work not only has direct implications for the design and operation of future urban public transport systems (e.g., more precise bus scheduling, improve service to public transport users), but also for urban planning (e.g., for transit oriented urban development) and travel forecasting.

This study is the first to employ an epidemiological framework to evaluate the ‘fit-for-purpose’ of ICD-10-AM external cause of injury codes, ambulance and hospital clinical documentation for injury surveillance. Importantly, this thesis develops an evidence-based platform to guide future improvements in routine data collections used to inform the design of effective injury prevention strategies. Quantification of the impact of ambulance clinical records on the overall information quality of Queensland hospital morbidity data collections for injury causal information is a unique and notable contribution of this study.

La produzione di energia a partire da biomassa rappresenta una delle importanti alternative all’utilizzo di combustibili fossili, al fine di ridurne il consumo globale e di limitare le emissioni di gas a effetto serra. In questo contesto il processo Biofine si classifica come una delle emergenti tecnologie per la produzione di acido levulinico a partire da biomassa lignocellulosica di seconda generazione. L’acido levulinico può essere successivamente idrogenato a γ-valerolactone, a sua volta alla base di processi di produzione di biocombustibili. Tuttavia, recenti studi hanno rivelato un aumento del numero di incidenti rilevanti nella catena di produzione di bioenergia maggiore del trend di produzione di energia stessa. Ciò, quindi, rende necessarie procedure di valutazione del rischio anche per questo tipo di impianti. Al fine di effettuare la valutazione del rischio dei due processi oggetto di studio, è stata adottata la metodologia ARAMIS, la quale si basa principalmente sulla costruzione di diagrammi bow-tie per l’identificazione dei Major Accident Hazards, e matrici del rischio, per l’identificazione dei Relevant Accident Scenarios. L’innovazione portata da questo studio riguarda la quantificazione delle consequenze, resa possibile attraverso il calcolo di distanze di danno, che permette di identificare scenari incidentali più realistici di quelli altrimenti previsti dalla ARAMIS.

This doctoral thesis focuses on the approaches currently in use to develop new vital valve devices: heart valve tissue engineering and heart valve tissue guided regeneration. Indeed, it investigates some aspects able to modify the preservation of the implanted construct, i.e. the immunogenic properties of the extracellular matrix and the biological entity of stem cell populations in heart valve leaflets. The porcine semilunar valves, as the bovine pericardial derived ones, find large clinical usage for the substitution of dysfunctional valves. The rational of their employ is the high morphological and functional analogies with the native human ones and hence they can provide a valid extracellular matrix (ECM) for bioengineered prosthesis. Once eliminated the xenogeneic cell component through a decellularizing treatment, it is possible to benefit from the extracellular fiber mesh as a template for repopulation with human stem cells in order to develop autologous-like replacements. So far the influence on cellular attachment exerted by the anisotropic ECM distribution in fibrosa and ventricularis has never been investigated. Porcine pulmonary leaflets were decellularized with a Triton X110/Sodium Cholate-based protocol. The absence of alpha-Gal epitopes, highly immunogenic for the human species, has been demonstrated through a double fluorescence technique using an isolectin and a specific monoclonal antibody. After treatment with fetal bovine serum and fibronectin to increase adhesion, human bone marrow mesenchymal stem cells have been seeded either onto the ventricularis or fibrosa and statically maintained in culture for 30 days. Not only ventricularis is able to consent a higher attachment, but also increased spreading and early cell differentiation in the cusp stroma. The interaction between the same cells and decellularized human leaflets has been further evaluated by ventricularis seeding. The homologous combination favors proliferation, with decrease of apoptotic events and enhanced cell maturation level, so that it is possible to appreciate the expression of typical mature smooth muscle markers in ventricularis. ECM performs consequently an essential role for cell integration by furnishing specific signals for the acquisition of the correct valvular phenotype. A further hypothesis for the achievement of new vital substitutes is likely represented by tissue guided regeneration operated in vivo by ECM in respect to the recipient's cells. Decellularized aortic roots were used to mimic a reconstruction of the RVOT in minipigs and performances were followed echocardiographically for 12-14 months, revealing a progressive function improvement. A continuous engraftment is observed in the explanted specimens, even if the layers less exposed to blood circulation are still devoid of cells. The observations on primary cultures of the vessel and leaflet tissues lead to hypothesize the contribution of two main phenotypes to repopulation: pulmonary artery smooth muscle cells and mobilized mesenchymal stem cells. Particularly important is the preservation of the bioconstruct once the function has been rescued, but at the same time it is essential to understand the key cell effectors participating to valve dysfunction in order to prevent the phenomenon. Scarcely studied, possible stem cell populations in the valves could be involved in the homeostatic tissue remodeling or in adverse pathophysiological events. Aortic roots (n=27) with related mitral leaflets (n=27) have been classified in 5 groups depending on allograft donor's age (10- 60 years). In addition, aortic and mitral pathological leaflets (n=10) were harvested during valve replacements. Cusp cryosections have been analysed through classical histology, Mallory's trichrome, Von Kossa, Oil Red O and immunohistochemistry for differentiated, inflammatory, calcifying, different lineagederived stem cell markers. With aging the typical observed feature is the accumulation of lipids, both as small droplets in the subventricularis layer or as cholesterol crystals in the fibrosa. Calcifications appear rare. Stem cell epitopes are highly expressed in the leaflet to reveal a specific spatial distribution. Primary cultures obtained from these specimens showed a phenotype and a differentiation potentiality, suggestive of mesenchymal stem progenitors without apparent osteogenic induction. Further studies are needed to better understand the contribution of these cells in heart valve calcification. In conclusion, both proposed modalities for the achievement of new valve substitutes appear valid instruments for the aim, because relying on fully decellularized alpha-Gal negative matrices permissive of cell-repopulation. Future efforts and application of new biomimetic strategies would surely implement the encouraging results here demonstrated. New insights on stem cell valvular biology have been proposed here for the first time with important confirmations on the continuous cell remodeling interesting the cusp tissue.<br>Tale tesi di dottorato si focalizza sugli approcci correntemente in uso per lo sviluppo di nuovi sostituti valvolari con caratteristiche di vitalità: l’ingegneria tissutale valvolare e la rigenerazione valvolare tessuto-guidata. Inoltre, mette in luce alcuni aspetti in grado di influenzare la preservazione dell’impianto, come le proprietà immunogeniche della matrice extracellulare e l’entità biologica delle popolazioni staminali nei lembi valvolari cardiaci. Le valvole semilunari porcine, come quelle di pericardio bovino, trovano largo impiego per la sostituzione di valvole non più funzionanti. Il razionale per l’impiego di esse è l’alta analogia morfologica e funzionale con le valvole umane, che le rende fonte ottimale di matrice extracellulare (ECM) per protesi bioingegnerizzate. Una volta rimossa la componente cellulare xenogenica attraverso un trattamento decellularizzante, è possibile beneficiare della trama di fibre extracellulari come supporto tridimensionale per il ripopolamento con cellule staminali umane in modo da sviluppare dei sostituti simil-autologhi. Finora, non è mai stata indagata l’influenza esercitata dalla distribuzione anisotropica delle fibre dell’ECM in ventricularis e fibrosa sull’adesione cellulare. Lembi polmonari porcini sono stati sottoposti a decellularizzazione mediante un trattamento a base di TritonX100/ Sodio Colato. L’assenza dell’epitopo alpha-Gal, altamente immunogeno per l’uomo, è stata dimostrata attraverso una doppia tecnica di fluorescenza con un’isolectina e uno specifico anticorpo monoclonale. Dopo condizionamento con siero bovino fetale e fibronectina per aumentare l’attaccamento, le cellule staminali mesenchimali del midollo osseo sono state seminate sia su ventricularis o fibrosa e mantenute staticamente in vitro per 30 giorni. Non solo la ventricularis favorisce una maggiore adesione, ma anche un’aumentata diffusione all’interno dello stroma e un precoce differenziamento cellulare nei citotipi valvolari. Successivamente, è stata valutata l’interazione fra cellule umane e matrice decellularizzata umana, effettuando ancora una volta la semina su ventricularis. La combinazione omologa permette di ottenere una maggiore proliferazione , una diminuzione degli eventi apoptotici e un maggiore grado di maturazione del costrutto, tanto da apprezzare l’espressione di proteine tipiche del muscolo liscio proprio nella tonaca di pertinenza, la vetricularis. L’ECM effettua quindi un ruolo essenziale per l’integrazione cellulare fornendo segnali specifici per l’acquisizione del fenotipo valvolare corretto. Un’ulteriore ipotesi per l’ottenimento di nuovi sostituti vitali può essere rappresentata dalla rigenerazione tessuto-guidata, favorita in vivo dall’ECM in grado di esercitare un ruolo attrattivo e di guida per le cellule dell’ospite. Radici aortiche decellularizzate sono state impiantate in minipig per mimare la ricostruzione del tratto di efflusso del ventricolo destro e la funzionalità valvolare è stata monitorata ecocardiograficamente per 12-14 mesi, rivelando un progressivo miglioramento della funzione. Un ripopolamento continuo può essere osservato attraverso le analisi ex vivo negli espianti, anche se le regioni meno esposte al flusso sanguigno sono ancora prive di cellule. Le osservazioni sulle colture primarie dei tessuti parietale vascolare e del lembo portano a ipotizzare il contributo di due maggiori fenotipi al ripopolamento: cellule muscolari lisce dell’arteria polmonare e cellule staminali mesenchimali mobilizzate. Particolarmente importante non è solo la creazione, ma anche il mantenimento del biocostrutto: una volta che si è ripristinata la funzione, perciò, è essenziale comprendere i reali effettori cellulari coinvolti nella disfunzione valvolare per prevenire il fenomeno già nelle prime fasi. Scarsamente studiate, eventuali popolazioni cellulari staminali a livello valvolare potrebbero partecipare al rimodellamento tissutale omeostatico, ma anche a eventi patologici avversi. Homograft valvolari aortici (n=27) con annesso lembo anteriore della mitrale (n=27) sono stati classificati in 5 gruppi a seconda della fascia d’età di appartenenza del donatore (10-60 aa). In aggiunta, sono stati esaminati anche lembi aortici e mitrali (n=10), rimossi in sede di sostituzione valvolare. Criosezioni di tessuto cuspidale sono state sottoposte all’analisi istologica classica, istochimica (von Kossa, Oil Red O) e immunoistochimica per marker di cellule differenziate, citotipi calcificanti precoci, flogosi e cellule staminali di vari lineages. All’aumentare dell’età, il tipico aspetto osservato é l’accumulo di lipidi, sottoforma di piccole gocce nella tonaca sottoventricolare o depositi di colesterolo nella fibrosa. Il riscontro di calcificazioni è raro. Gli epitopi delle cellule staminali sono espressi nel lembo valvolare a rivelare una precisa distribuzione spaziale. Le colture primarie ottenute da questi campioni mostrano un fenotipo e proprietà di transdifferenziazione, caratteristiche di un possibile progenitore mesenchimale, senza tuttavia apparente induzione all’osteogenesi. Ulteriori studi sono necessari per meglio comprendere il coinvolgimento di tali cellule nella patologia valvolare. Concludendo, entrambe le modalità considerate per la realizzazione di un nuovo sostituto vitale sembrano validi strumenti a raggiungere lo scopo, basandosi su matrici decellularizzate prive di alpha-gal e prontamente colonizzabili. Esperimenti futuri e l’applicazione di nuove strategie biomimetiche potranno sicuramente migliorare i risultati incoraggianti finora ottenuti. Sono state inoltre proposte per la prima volta nuove informazioni sulla biologia cellulare staminale della valvola cardiaca con importanti conferme sul processo di continuo rimodellamento cellulare a carico del tessuto cuspidale.

Le thé est la deuxième boisson la plus consommée au monde, dépassé uniquement par l'eau, en raison de ses bienfaits pour la santé. Cependant, en raison des pratiques de culture en monoculture, l'utilisation de pesticides pendant la culture du thé est courante. Au fil du temps, le nombre de pesticides utilisés a augmenté et, pour protéger la santé des consommateurs, de nombreux pays et régions ont établi des limites maximales de résidus de pesticides pour une variété d'aliments et de boissons, y compris le thé. Par exemple, l'Union européenne (UE) a fixé les limites maximales de résidus (LMR) pour plus de 480 pesticides et leurs métabolites dans les produits à base de thé. Par conséquent, le développement de méthodes analytiques pour les résidus multi-pesticides dans le thé est un défi, car le thé est un produit complexe avec de nombreux composés qui peuvent interférer avec les résultats, tels que les polyphénols, les pigments et la caféine. Le but de cette thèse est de développer des méthodes robustes et robustes avec une sensibilité, une exactitude et une précision élevées pour répondre aux réglementations de l'UE pour la détermination simultanée de 400 résidus de pesticides dans les produits de thé en utilisant des chromatographies liquides et gazeuses ultra performantes couplées à la spectroscopie de masse en tandem (UPLC- MS/MS et GC-MS/MS, respectivement). La première partie de la thèse portait sur l'élimination des effets de matrice dans les feuilles de thé vertes en combinant l'extraction QuEChERS (rapide, facile, bon marché, efficace, robuste et sûre) et le nettoyage en mode mixte SPE (extraction en phase solide). Une cartouche SPE C18 couplée à SPE GCB / PSA s'est avérée être la méthode de nettoyage la plus efficace et a permis de quantifier 225 résidus de pesticides, sur la base des courbes d'étalonnage des solvants (154 résidus utilisant UPLC-MS/MS et 71 résidus utilisant GC-MS/MS). Les méthodes analytiques ont été entièrement validées conformément au SANTE/11945/2015 (UE). Les LOQ pour la plupart des pesticides (386/400 ou 96,5%) étaient inférieures à 10 &#956;g / kg, c'est-à-dire inférieures à la LMR de l'UE (5-70 mg / kg). Dans la deuxième partie, les effets matriciels de 400 résidus de pesticides ont été étudiés et améliorés pour l'analyse de différents types de thés (blancs, verts, oolongs et noirs). Les résultats ont montré que la combinaison de l'extraction QuEChERS et du nettoyage SPE en mode mixte et après la réduction du volume d'injection s'est avérée être la procédure la plus efficace pour surmonter les effets de matrice. Plus de 190 pesticides (&gt; 95% des 200) ont eu l'effet matriciel dans une fourchette de ± 20% pour UPLC-MS/MS. Par conséquent, ils peuvent être quantifiés à l'aide de courbes d'étalonnage de solvant. D'un autre côté, des courbes d'étalonnage adaptées à la matrice doivent être utilisées pour surmonter les effets de matrice pour GC-MS/MS. De plus, nous avons reconnu que les effets de matrice dans GC-MS/MS n'étaient pas seulement une amélioration du signal mais aussi une suppression. Enfin, dans la troisième partie de ces travaux, la méthode établie a été appliquée avec succès à la détermination des résidus multi-pesticides dans 106 échantillons de thé. Au total, 26 échantillons de thé (24,5%) contenaient au moins une violation de pesticides, avec 43 violations de résidus de pesticides. Les pesticides les plus fréquemment détectés étaient les néonicotinoïdes, les pyréthrinoïdes synthétiques et les fongicides triazolés. En termes d'origine dans cette étude, Taïwan avait les échantillons les plus contaminés par les pesticides avec 83,3%, suivi par la Chine (73,7%), le Vietnam (64,7%) et l'Inde (Darjeeling) (55,0%). Les résultats ont montré que les échantillons dépassant les réglementations de l'UE en matière de LMR étaient toujours élevés à 24,5%. Par conséquent, les évaluations des multiples résidus de pesticides dans le thé doivent être poursuivies<br>Tea is the second most-consumed beverage in the world, surpassed only by water, due to its health benefits. However, because of monoculture cultivation practices, the use of pesticides during tea cultivation is common. Over time, the number of pesticides used has increased, and, to protect consumers' health, many countries and regions have established maximum residue limits of pesticides for a variety of foods and beverages, including tea. For instance, the European Union (EU) has set the maximum residue limits (MRLs) for more than 480 pesticides and their metabolites in tea products. Therefore, the development of analytical methods for multi-pesticide residues in tea is a challenge, because tea is a complex commodity with many compounds that can interfere with results, such as polyphenols, pigments, and caffeine. The aim of this thesis is to develop rugged and robust methods with high sensitivity, accuracy, and precision to meet the EU regulations for simultaneous determination of 400 pesticide residues in tea products using ultra performance liquid and gas chromatographies coupled to tandem mass spectroscopy (UPLCMS/ MS and GC-MS/MS, respectively). The first part of thesis focused on elimination of matrix effects in green tealeaves by combining QuEChERS (quick, easy, cheap, effective, rugged, and safe) extraction and mixed-mode SPE (solid phase extraction) clean-up. A C18 SPE cartridge paired with SPE GCB/PSA proved to be the most effective clean-up method and enabled 225 pesticide residues to be quantified, based on solvent calibration curves (154 residues using UPLCMS/ MS and 71 residues using GC-MS/MS). The analytical methods were validated fully in accordance with the SANTE/11945/2015 (EU). LOQs for most pesticides (386/400 or 96.5%) were below 10 &#956;g/kg, i.e., less than the EU MRL (5-70 mg/kg). In the second part, matrix effects for 400 pesticide residues were investigated and improved for the analysis of different types of teas (white, green, oolong and black ones). Results showed that combining QuEChERS extraction and mixed-mode SPE clean-up, and following the reduction of the injection volume were found to be the most effective procedure to overcome matrix effects. More than 190 pesticides (&gt; 95% of the 200 ones) had the matrix effect within the range of ± 20% for UPLC-MS/MS. Therefore, they can be quantified using solvent calibration curves. On the other hand, matrix-matched calibration curves should be used to overcome matrix effects for GC-MS/MS. Moreover, we recognized that matrix effects in GC-MS/MS were not only signal enhancement but also suppression. Finally, in the third part of this work, the established method was successfully applied to the determination of multi-pesticide residues in 106 tea samples. In total, 26 tea samples (24.5%) were containing at least one pesticide violation, with 43 pesticide residue violations. The most frequently detected pesticides were neonicotinoids, synthetic pyrethroids, and triazole fungicides. In terms of origin in this study, Taiwan had the most pesticide-contaminated samples with 83.3%, following by China (73.7%), Vietnam (64.7%), and India (Darjeeling) (55.0%). The results showed that samples exceeding EU MRLs regulations were still high with 24.5%. Therefore, assessments of multipesticide residues in tea need to be continued

Livestock production is one of the largest and fastest growing agricultural subsectors in the world, contributing almost 40 % of agricultural GDP. The present global trend of livestock intensification affects water demand, resulting in local alterations of hydrological features. This study quantifies water requirements for livestock production and links them to local changes in water-related ecosystem services (WRES). It examines the present situation in Uruguayan beef production as it undergoes an intensification process by analyzing three systems: traditional extensive production, mixed systems and intensive production.Mixed production systems are most water-effective in terms of water requirement. However, extensive production has the least negative long-term effects on the studied WRESs: soil productivity, erosion control and grass, crop and livestock production. Intensive systems contribute the largest negative impacts on soil productivity due to management practices, e.g. continuous cropping and fertilizer application (which is also costly). However, intensification also increases short-term positive effects on crop and livestock production. Mixed systems offer opportunities to increase livestock production, water use efficiency and mitigate soil degradation. However, a threshold value to avoid an undesirable regime shift in soil productivity remains to be determined.This framework developed herein, maps and quantifies local ecosystem effects associated with livestock production from a water perspective. It required simplifications and thus needs further improvements to be of practical use as a tool for best management practices of livestock production. Such tools can assist decision-makers in Uruguay on how to manage multiple ecosystem services in an agricultural landscape and enhance the provision of desired ES without degrading other key ecosystem processes, goods and services. For sustainable production of livestock in Uruguay, one should consider if the tradeoff between long-term negative impacts of intensification on regulating services is justifiable in the context of the short-term economic gains in provisioning services.

The rapid and accurate identification of infectious agents and the early introduction of an appropriate antimicrobial therapy is of crucial importance in the clinical practice, mainly in serious infections such as blood stream infection (BSI), which may progress to sepsis /septic shock with high morbidity and mortality rate. The delay in an appropriate antibacterial treatment has led to increases non only in morbidity and mortality rates as well as in hospitalization time, in nosocomial-acquired infections and make it challenging to implement the back-end approach of the antimicrobial stewardship program, which has shown rewarding results in patient management and the fight against the antimicrobial resistance. Moreover, the emergence of resistant and multidrug- or even pandrug-resistant pathogens, especially the carbapenemaseproducing Enterobacteriacea (CPE) such as Klebsiella pneumoniae carbapenemase-producing (KPC) K. pneumoniae strains, is increasingly reported worldwide and is becoming an important issue in health care systems. Therefore, there is an urgent requirement to provide timely information on infection status, which allow an optimal patient’s management and treatment. Clinical microbiology has developed continuously and there has been a constant search for new techniques and diagnostic methods with the aim to optimize the identification of pathogens, their antimicrobial susceptibility profile and improve the patient’s management and treatment reducing the social and economic impact on the health care system. Molecular methods seem to be an appropriate choice, they are widely used in the diagnosis of BSIs, alongside to the conventional culture-based methods. However, these techniques involve significantly increased cost and technical complexity, both of which are likely to hamper their adoption in the laboratory routine in the clinical setting. In this paper are reported two novel technologies in clinical microbiology focused on the rapid diagnosis of BSIs directly from clinical material in order to reduce the Turn Around Time (TAT) for bacterial identification and improve patient’s management and treatment. In the first part of this study has been evaluated the performance of a new molecular assay, the beacon-based fluorescent in situ hybridization (bbFISH) in comparison to the conventional growthbased phenotypic identification of bacteria from positive blood culture vials in febrile patients. It has been also examined the bbFISH and the conventional identification assay’s total turnaround time (TAT) performance. Considering the good sensitivity and specificity of the hemoFISH® assays as well as the significant time saving in obtaining the final results (p-value 0.0001), the introduction of this assay could be reliable in the microbiology laboratories, supplementing traditional approaches, speeding up the diagnosis of bloodstream infections and identifying the majority of most important sepsis pathogens also in the case of polymicrobial infection. This assay has the potential to provide timely and cost effective information on infection status, thus allowing clinicians to make more informed decisions on appropriate antibiotic therapy at an earlier stage than is possible with culture-based approaches. hemoFISH® provides a same-day identification of the majority of microorganisms thus the turnaround time is considerably lower than microbiological culture. The second part of the present work is instead dedicated to another novel proteomic assay, the Matrix-Assisted Laser Desorption Ionization-Time of flight Mass Spectrometry (MALDI-TOF MS). The application of this technology for the identification of pathogens in blood culture (BCs) has been significantly improved the time required to obtain results with an excellent sensitivity and specificity. Results were available on average after only a few minutes, representing a significant difference in comparison with the time required for the results of conventional methods. Moreover the accuracy of this identification method seems to be considerable, with higher performance and cost effectiveness comparing with traditional methods. In addition, has been evaluated the applicability of this technology in order to predict susceptibility or resistance profiles against antibiotics, such as the carbapenems, on BCs derived directly from patients with Gram-negative bacteremia, in particular with carbapenemaseproducing Klebsiella pneumonia strains. The results of the present study have demonstrated that MALDI-TOF MS technology is a relevant tool also for the detection of antibiotic resistance. The enzymatic hydrolysis of the amide bond into the β-lactam ring of carbapenems, like ertapenem, by β-lactamases enzymes could be monitored by MALDI-TOF MS assay; hydrolyzed and nonhydrolyzed substances differ in their molecular weights and these differences can be detected by this assay within 1 to 3 hours with a great accuracy and high sensitivity. Compared to other alternative approaches like molecular genetics techniques this method is cost effective and easy to perform. The rapid detection of clinically important β-lactamases in routine diagnostic laboratories may be crucial for initial antibiotic therapy as well as for the prevention of the speed of β- lactamase-producing bacteria in health care setting. Combined with a proactive antimicrobial stewardship program of carbapenems this assay can improve patients outcomes. In order to provide even more timely information on infectious status, which allow an optimal patient’s management and treatment, always in the second part of this study has been also demonstrated the reliability of a rapid antibiotic susceptibility test (AST) performed directly from BCs flagged as positive to Gram-negative rods. The comparison of the antibiotic susceptibility data obtained by the standard method and by the direct method has showed a concordance of 100% while these data were obtained in a shorter time by the direct method than the traditional culture-based susceptibility testing. Thus, this methodological approach has great potential to become a routine method in any clinical microbiology laboratory while both this one method and MALDI-TOF MS technology can likely modified the workflow of BCs providing a same-day final report available to clinicians and hospital epidemiologists.

The work described in this thesis was undertaken to investigate physicochemical, biochemical and physiological factors contributing towards the rarity of kidney stone disease in the South African black population. Healthy, age-matched male subjects from the black and white population groups were recruited for this purpose. In several of the studies, subjects followed a standardized diet and were required to provide 24 hour urine collections. These were analyzed for sodium, potassium, calcium, oxalate, uric acid, citrate, chloride, magnesium, phosphate, sulphate and creatinine using standard laboratory techniques. Urine composition values were used as input data for the calculation of relative supersaturation (RS) values for calcium oxalate (CaOx), calcium phosphate (CaP, or brushite) and uric acid (U A) using the computer programme EQUIL and for the calculation of the Tiselius Risk Index (TRI). CaOx crystallization experiments were performed. These included CaOx metastable limit (MSL) and BONN Risk Index (BRI) determinations, particle formation kinetics, 14 C-oxalate cry stal deposition kinetics and CaOx crystal aggregation and nucleation inhibition. Crystallizati on experiments were also supplemented with scanning electron microscopy (SEM) and zeta potential measurements. Urine compositions, crystallization data and physicochemical risk indices were analyzed statisitically using ANOV A. Several different investigations were undertaken. These included crystallization experiments involving urinary macromolecules from both race groups, crystal matrix extract isolation (with osteopon tin as its major component) from both race groups and its testing for inhibitory capacity in ultrafiltered urine from both race groups. Similar crystallization experiments were conducted with commercially available osteopontin. In addition, a comprehensive trial was conducted in which the ingestion of three sodium salts (sodium chloride, sodium bicarbonate and sodium citrate) was investigated for their effects on urinary risk factors of CaOx stone formation and for their effects on bone turnover markers in urine and in serum. For the biochemical isolation of crystal matrix extract, COD-CME was precipitated in urine from both black and white subjects. The proteins included in COD-CME were detected using sodium dodecyl sulfate polyacryalamide gel electrophoresis (SDS-PAGE). Western Blotting was used for semi-quantitative analysis of OPN v For the trial involving different sodium salts, four experimental protocols were investigated. The four protocols included low NaCl (3 g/day), high NaCl (12 g/day), sodium bicarbonate (6 g/day) and sodium citrate in the form of Citro-Soda (16 g/day). A Latin Square Design was followed for the random assignment of participants to sequences of protocols. The studies on macromolecules showed that those in the urine of black subjects were more potent inhibitors of CaOx crystal deposition and aggregation than those in the urine of white subjects. Isolation and characterization of the crystal matrix extract in COD crystals confirmed that osteopontin is the main intracrystalline protein in the extract. The crystallization experiments performed on the crystal matrix extract isolated from the urine of both race groups demonstrated that the extract isolated from the urine of black subjects was a superior inhibitor of CaOx deposition, growth and aggregation. Crystallization studies performed on commercially available osteopontin in urine from black and white subjects showed that this protein is a more effective inhibitor of CaOx crystal deposition, growth and nucleation in the urine from the former group compared to that from the latter. The studies on supplemental sodium salts demonstrated that the two race groups respond differently to lithogenic and anti-lithogenic dietary challenges. High NaCl protocol resulted in a favourable and counter-intuitive significant decrease in free unbound calcium in samples from black subjects whereas no such change was observed in white subjects. Supplemental sodium bicarbonate and sodium citrate induced favourable decreases in urinary total calcium, urinary ionized calcium, BRI, RS of CaOx, RS of uric acid and RS of brushite, and a favourable increase in urinary citrate and pH in both groups. Interestingly and more importantly, these factors were more prominent in samples from black subjects than those from white subjects. Bone turnover measurements showed that urinary deoxypyridinoline (DPD) levels were lower while serum osteocalcin (OC) levels were higher in blacks than in whites at baseline, but these differences were not statistically significant. Smaller increases in urinary DPD levels after high aCl and sodium bicarbonate and corresponding bigger increases in serum osteocalcin (OC) levels after these protocols (and sodium citrate) in black subjects than in white subjects indicate less bone resorption and higher bone formation, respectively, in the former group. Vl The results presented in this thesis have provided convincing evidence that in the context of CaOx kidney stone formation, several physicochemical, biochemical and physiological factors are different in black and white South African subjects and that these factors are more effective in the former group with respect to providing protective mechanisms against CaOx kidney stone formation

Lors de la migration cellulaire, les complexes d’adhérence contrôlent la production de force et l’adaptation aux propriétés mécaniques de l’environnement. Le but de ce projet était de comprendre les mécanismes moléculaires par lesquels ces complexes contrôlent la force produite par l’assemblage du cytosquelette d’actine et codent une information mécanique en réaction biochimique. La première étude montre que les protéines taline, RIAM et VASP s’assemblent à la surface d’une membrane pour stimuler l’assemblage de l’actine par un nouveau mécanisme. La deuxième partie est basée sur la reconstitution de la machinerie mécano-sensible des complexes d’adhérence avec des protéines pures sur une surface micro-imprimée observée en microscopie. L’étude montre que la protéine étirable taline échange son partenaire RIAM contre la vinculine en réponse à la force du cytosquelette. La taline code donc l’information mécanique en recrutant les partenaires qui correspondent à son degré d’étirement<br>During cell migration, adhesion complexes control the production of force and the adaptation to the mechanical properties of the environment. The aim of this project was to understand the molecular mechanisms by which these complexes control the force produced by actin assembly and encode mechanical information into biochemical reactions. The first study shows that the proteins talin, RIAM and VASP assemble on the surface of a membrane to stimulate actin assembly by a novel. mechanism The second part is based on the reconstitution of the mechanosensitive machinery of the adhesion complexes with pure proteins on a micropatterned surface observed in microscopy. The study reveals that the stretchable protein talin exchanges its partner RIAM for vinculin in response to the force transmitted by the actin cytoskeleton. Talin thus codes mechanical information by recruiting partners that correspond to its degree of stretch

The first and foremost goal of the present work was to develop novel theoretical and computational methods and use state-of-the-art techniques in electronic structure theory to interpret a specific set of physical problems mainly related, but not limited to, materials science. Our guiding principle was to relate information obtained from scattering experiments with the numerical solution of the multichannel dynamics of many-body systems, shedding light on the origin of electronic and optical properties of a variety of systems. The general approach adopted in this thesis was not to present separate chapters for theory, rather we introduced methods along with the experiments. In particular, we focused on the modeling of both ground and excited states of materials, on vibrational, core and valence electron spectroscopy of condensed matter systems using computational methods at different level of accuracy and complexity to interpret a number of experimental data. While these methods have been devised for this scope, their applicability, notably the treatment of the continuum states through multichannel scattering formalism, is totally general and can be applied to describe several different experiments, performed with a variety of apparently distant techniques. In particular, the Fano--Fesbach discrete-continuum interaction provides a common framework suitable to this task. Within this scheme, thus, the calculation of the spectral lineshapes measured by XPS, Auger, NEXAFS, and EEL spectroscopy can be reconciled on the same theoretical grounds with the investigation of the properties of ultra-cold Fermi gases at unitarity, or of the electronic capture and decay rate in ultra-hot plasma found in stellar environments or, finally, with the study of the epitaxial growth of nanostructured materials. Crossing the borders between several computational, theoretical and experimental techniques, this thesis should be of interest to a broad community, including those interested in aspects of atomic and molecular physics, electronic structure calculations, experimental and theoretical spectroscopy, astrophysics and scattering theorists in a broad sense.

The first and foremost goal of the present work was to develop novel theoretical and computational methods and use state-of-the-art techniques in electronic structure theory to interpret a specific set of physical problems mainly related, but not limited to, materials science. Our guiding principle was to relate information obtained from scattering experiments with the numerical solution of the multichannel dynamics of many-body systems, shedding light on the origin of electronic and optical properties of a variety of systems. The general approach adopted in this thesis was not to present separate chapters for theory, rather we introduced methods along with the experiments. In particular, we focused on the modeling of both ground and excited states of materials, on vibrational, core and valence electron spectroscopy of condensed matter systems using computational methods at different level of accuracy and complexity to interpret a number of experimental data. While these methods have been devised for this scope, their applicability, notably the treatment of the continuum states through multichannel scattering formalism, is totally general and can be applied to describe several different experiments, performed with a variety of apparently distant techniques. In particular, the Fano--Fesbach discrete-continuum interaction provides a common framework suitable to this task. Within this scheme, thus, the calculation of the spectral lineshapes measured by XPS, Auger, NEXAFS, and EEL spectroscopy can be reconciled on the same theoretical grounds with the investigation of the properties of ultra-cold Fermi gases at unitarity, or of the electronic capture and decay rate in ultra-hot plasma found in stellar environments or, finally, with the study of the epitaxial growth of nanostructured materials. Crossing the borders between several computational, theoretical and experimental techniques, this thesis should be of interest to a broad community, including those interested in aspects of atomic and molecular physics, electronic structure calculations, experimental and theoretical spectroscopy, astrophysics and scattering theorists in a broad sense.

This thesis deals with the theoretical and computational modelling of materials by using a variety of ab-initio approaches to accurately predict the properties of realistic structures. A number of known and novel carbon-based materials are studied, exploiting the unique versatility of carbon to bind into several bonding configurations, with the aim of tailoring their electronic and mechanical characteristics. In this regard, the methods used to carry out electronic structure simulations depend on the system size: from the Dirac-Hartree-Fock approach to model molecular properties, to Density Functional Theory used for periodic solids, such as diamond and graphene-related materials composed by a few to some hundred of atoms, to Density Functional Tight Binding or plane Tight Binding to study nanowires or Beltrami pseudospheres, which are composed by some hundreds to a few millions atoms. The details of these methods are introduced in the chapters where they are used. The criterion used to present these concepts is to organize the chapters, with the exception of the last one, according to the increasing dimension of the systems. More in details, the first chapter uses the Dirac-Hartree-Fock approach to simulate atoms and molecules, such bromotrifluoromethane; the second chapter deals with periodic systems characterized by unit cells with a relatively small number of atoms, such as diamond and graphite; the third one discusses graphene and graphene-related materials with lower density; the fourth one present a new computational and experimental model of silicon carbide nanowires coated with silicon dioxide shell; the fifth chapter is focused on the study of sp2-hybridized carbon atoms, arranged on a Beltrami surface. The latter topic spans different research fields such as geometrical topology, physics and mechanical engineering. Finally, the last chapter is dedicated to an on going work which deals with the Non-Adiabatic Molecular Dynamics simulation of amorphous silica samples where we couple the nuclear dynamic of the system to the electronic structure.

This thesis deals with the theoretical and computational modelling of materials by using a variety of ab-initio approaches to accurately predict the properties of realistic structures. A number of known and novel carbon-based materials are studied, exploiting the unique versatility of carbon to bind into several bonding configurations, with the aim of tailoring their electronic and mechanical characteristics. In this regard, the methods used to carry out electronic structure simulations depend on the system size: from the Dirac-Hartree-Fock approach to model molecular properties, to Density Functional Theory used for periodic solids, such as diamond and graphene-related materials composed by a few to some hundred of atoms, to Density Functional Tight Binding or plane Tight Binding to study nanowires or Beltrami pseudospheres, which are composed by some hundreds to a few millions atoms. The details of these methods are introduced in the chapters where they are used. The criterion used to present these concepts is to organize the chapters, with the exception of the last one, according to the increasing dimension of the systems. More in details, the first chapter uses the Dirac-Hartree-Fock approach to simulate atoms and molecules, such bromotrifluoromethane; the second chapter deals with periodic systems characterized by unit cells with a relatively small number of atoms, such as diamond and graphite; the third one discusses graphene and graphene-related materials with lower density; the fourth one present a new computational and experimental model of silicon carbide nanowires coated with silicon dioxide shell; the fifth chapter is focused on the study of sp2-hybridized carbon atoms, arranged on a Beltrami surface. The latter topic spans different research fields such as geometrical topology, physics and mechanical engineering. Finally, the last chapter is dedicated to an on going work which deals with the Non-Adiabatic Molecular Dynamics simulation of amorphous silica samples where we couple the nuclear dynamic of the system to the electronic structure.

Molecular nanomagnets have attracted the attention of the scientific community since they represent model systems to study many quantum phenomena, such as quantum entanglement and decoherence. They are also interesting for their envisaged technological applications, including magnetic refrigeration, high-density information storage and quantum information processing. Driven by these motivations, the present Thesis focuses on the study of magnetic properties and spin dynamics of different classes of molecular nanomagnets, with the purpose to understand their quantum behavior and their potential applications in future technologies. Magnetic frustration is at the origin of many exotic phenomena in matter. Here frustration-induced properties are studied and analyzed for the clusters Ni7, Fe7, Fe6 and Mn6, together with their effects on phonon-induced relaxation dynamics. Indeed, the comprehension of relaxation mechanisms is crucial in order to address the implementation of these systems in the fields of quantum information processing or information storage. Molecular nanomagnets are also promising materials for very-low-temperature magnetic refrigeration due to a potentially enhanced magnetocaloric effect. By explicitly considering Carnot refrigeration cycles, the theoretical recipe to design the best molecules for cryogenic magnetic refrigeration is identified. Another important class of molecular nanomagnets is that of antiferromagnetic rings. Of these, Cr7Ni is of great importance for applications in quantum information processing. The local spin density distribution in Cr7Ni is here obtained from 53Cr-NMR and confirmed by theoretical calculations. The origin of magnetic anisotropy of these rings grafted on surfaces is investigated using XMCD and theoretical calculations. Antiferromagnetic “purple” rings are completely characterized by the comparison of our calculations with inelastic neutron scattering and EPR data. The characterization of purple rings represents a first step in the description of a new family of “purple-green” entangled dimers, which represent model systems to study spin entanglement and its application in the field of quantum computation.

Although protein folding has been studied for decades many open issues still resist, and we yet lack a clear and general description of the mechanisms leading from the unfolded to the folded state. In particular, it is still under debate whether proteins fold through few well-defined pathways or trough a large multitude of independent ways. Answering these questions is made difficult by the fact that standard molecular dynamics (MD) simulations are very computationally expensive and often impracticable. Moreover, often even experimental techniques lack the necessary resolution to give a definitive answer. We will introduce and develope the Dominant Reaction Pathway (DRP), which is an approach that permits to efficiently study the thermally activated conformational dynamics of bio-molecules in atomistic detail. In particular, it can be used to characterize and portray the folding pathways of a protein once the unfolded and folded configurations are given. We firstly applied the DRP to a realistic protein studying the folding pathways of the Fip35 WW Domain, a 35 amino-acids long protein. Performing all atom simulations we were able to show that this small protein folds following only two pathways, defined by the order of formation of secondary structures. Notably, our results are compatible with ultra long MD simulations and consistent with the analysis of the experimental available data on the folding kinetics of the same system. Exploiting the efficiency of the DRP formalism, computing a folding trajectory of this protein only required about one hour on 48 CPU’s. We applied then our simulation scheme to a much more challenging task: performing an all-atom folding simulation of a 82 amino-acids long protein displaying a topological knot in its native conformation. We were able to portray the folding mechanism and to identify the essential key contacts leading to the proper formation of this knot. Interestingly, we showed that non native contacts, i.e., transient contacts formed during the folding of the protein but absent in its native state, can sensibly enhance the probability of correctly forming the knot.

This thesis studies problems concerning the dynamics and thermodynamics of manybody quantum systems. We start by introducing the necessary theoretical concepts and tools forming the basis of this manuscript. The research presented can be split in two parts. The first one deals with the dynamics of many-body quantum systems subject to environmental dissipative effects of various forms, while the second one studies topics of thermodynamics in many-body quantum systems. The first part of the research presented studies the effects of an environment inducing dissipation. We establish and study the adiabatic dynamics of free-fermion models subject to a local Lindblad bath and in the presence of a time-dependent Hamiltonian. The merit of these models is that they can be solved exactly, which thus can help us to study the interplay between non-adiabatic transitions and dissipation in many-body quantum systems. After the adiabatic evolution, we evaluate the excess energy (average value of the Hamiltonian) as a measure of the deviation from reaching the target final ground state. We find a robust evidence of the fact that an optimal working time for the quantum annealing protocol emerges as a result of the competition between the non-adiabatic effects and the dissipative processes. We compare these results with matrix product operator simulations of an Ising system and show that the phenomenology we found applies also for this more realistic case. We then proceed to a scenario in which the environment is not detrimental, but is on the contrary the driving force of the effects studied. We demonstrate that persistent currents can be induced in a quantum system in contact with a structured reservoir, without the need of any applied gauge field. The working principle of the mechanism leading to their presence is based on the extension to the many-body scenario of non-reciprocal Lindblad dynamics. Specifically, we consider an interacting spin/boson model in a ring-shaped one-dimensional lattice coupled to an external bath. By employing a combination of cluster mean-field, exact diagonalization and matrix product operator techniques, we show that solely dissipative effects suffice to engineer steady states with a persistent current that survives in the limit of large systems. We also verify the robustness of such current in the presence of additional dissipative or Hamiltonian perturbation terms. The second part studies many-body quantum systems with a focus on thermodynamics. First, we investigate a quantum battery made of N two-level systems, which is charged by an optical mode via an energy-conserving interaction. We quantify the fraction of energy stored in the battery that can be extracted in order to perform thermodynamic work. We first demonstrate that this fraction is highly reduced by the presence of correlations between the charger and the battery or between the two-level systems composing the battery. We then show that the correlation-induced suppression of extractable energy, however, can be mitigated by preparing the charger in a coherent optical state. We conclude by proving that the charger-battery system is asymptotically free of such locking correlations in the N ! 1 limit. And lastly, we study open questions within the theory of open quantum systems. The Markovian evolution of an open quantum system is characterized by a positive entropy production, while the global entropy gets redistributed between the system and the environmental degrees of freedom. Starting from these premises, we analyse the entropy variation of an open quantum system in terms of two distinct relations: the Clausius inequality, that provides an intrinsic bound for the entropy variation in terms of the heat absorbed by the system, and an extrinsic inequality, which instead relates the former to the corresponding entropy increment of the environment. By modeling the thermalization process with a Markovian collisional model, we compare and discuss the two bounds, showing that the latter is asymptotically saturated in the limit of large interaction time. In this regime not only the reduced density matrix of the system reaches an equilibrium configuration, but it also factorizes from the environmental degrees of freedom.

Two centuries after its first discovery, thermoelectricity, i.e. the phenomenon of direct conversion of thermal power into electrical power, has only recently reached the possibility of implementation in a vast number of practical applications. This breakthrough was undoubtedly determined by the advent and diffusion of numerical atomistic simulation techniques, allowing a quick survey of large classes of materials as possible candidates for the realization of active parts for thermoelectric generators. To this aim, two classes of materials have been essentially identified, (I) inorganic thermoelectrics, based on metal alloys, and (II) organic conductive polymers. The latter ones are well suited for implementation in a large-scale economy because of their superior mechanical properties, such as flexibility and low specific weight, as well as simpler synthesis process, as spin coating, and the possibility of tuning the electrical conductivity through chemical doping. Among the most common conducting polymers, polyethylenedi-oxythiophene (PEDOT), the subject of this Thesis work, has clearly emerged as one of the most promising thermoelectric material. Despite its wide use, however, an unanimous and well-established understanding of the link existing between the synthesis process and the corresponding final thermoelectric properties is still missing and it is thus an active field of scientific investigation. The present Thesis represents a part of this research stream, specifically aiming to shed a light on the role and effect of the combinations of most commonly used polymerizing reagents for PEDOT in determining the micromorphology and the resulting thermal and electrical transport properties. In this respect, the description and development of a new computational algorithm, based on a multiscale approach, is presented, combining a purely quantum description based on the Density Functional Theory (DFT) with Classical Molecular Dynamics (MD). The comparison of the results obtained by numerical simulation with the experimental data currently available demonstrates the effective possibility of including the chemical description of the synthesis process in the context of an MD simulation, and allows to demonstrate and quantify the impact of the combination of reagents used on (i) micromorphological properties, such as chain length distribution and crystallinity, (ii) thermal transport properties, in particular thermal conductivity, and (iii) carrier transport properties, mainly hole mobility and conductivity, estimated by Marcus' theory of electron transport.

Although protein folding has been studied for decades many open issues still resist, and we yet lack a clear and general description of the mechanisms leading from the unfolded to the folded state. In particular, it is still under debate whether proteins fold through few well-defined pathways or trough a large multitude of independent ways. Answering these questions is made difficult by the fact that standard molecular dynamics (MD) simulations are very computationally expensive and often impracticable. Moreover, often even experimental techniques lack the necessary resolution to give a definitive answer. We will introduce and develope the Dominant Reaction Pathway (DRP), which is an approach that permits to efficiently study the thermally activated conformational dynamics of bio-molecules in atomistic detail. In particular, it can be used to characterize and portray the folding pathways of a protein once the unfolded and folded configurations are given. We firstly applied the DRP to a realistic protein studying the folding pathways of the Fip35 WW Domain, a 35 amino-acids long protein. Performing all atom simulations we were able to show that this small protein folds following only two pathways, defined by the order of formation of secondary structures. Notably, our results are compatible with ultra long MD simulations and consistent with the analysis of the experimental available data on the folding kinetics of the same system. Exploiting the efficiency of the DRP formalism, computing a folding trajectory of this protein only required about one hour on 48 CPU’s. We applied then our simulation scheme to a much more challenging task: performing an all-atom folding simulation of a 82 amino-acids long protein displaying a topological knot in its native conformation. We were able to portray the folding mechanism and to identify the essential key contacts leading to the proper formation of this knot. Interestingly, we showed that non native contacts, i.e., transient contacts formed during the folding of the protein but absent in its native state, can sensibly enhance the probability of correctly forming the knot.

2014 - 2015<br>In the scientific community there is a great interest to explore new superconducting materials suitable for high field applications in order to meet the needs of industrial claims. In this framework, newly discovered Fe-Based Superconductors (IBSC) are a promising choice, especially due to their critical temperature intermediate between low and high Tc materials, as well as an extremely high upper critical field. The aim of this work has been the preparation and the study of physical properties of iron-chalcogenides superconducting samples, in particular polycrystalline FeSe and FeSeTe. The iron-chalcogenides family has been choosen mostly because of its interesting superconducting properties and also due to its simple crystalline structure and to the lack of poisonous elements in its composition. Opening a completely new research field at the ENEA CR Frascati, several routes of samples production have been carried out. I achieved part of the necessary know-how working also in other laboratories that have great experience on iron-based superconductors preparation, in particular the National Institute for Materials Science (NIMS) laboratories of Tsukuba in Japan, where I worked at the Nano Frontier Materials Group, under the leadership of Prof. Dr. Takano. I also had the chance to spend a brief period at the laboratories of CNR SPIN Genova and the Physics and Chemistry Departments at University of Genova, where I could meet researchers skilled in the production of iron-based samples. Most of the know-how was achieved by direct experience. Even if some of the routes for samples preparation did not brought to the expected results, some of these techniques gave interesting results, other routes deserve further optimization. Concerning the FeSe compound, two preparation processes have been implemented: the electrochemical deposition on iron substrate, and the solid state reactive synthesis. The former gave FeSe thin films containing the right tetragonal -phase, but the optimization of the superconducting properties in these samples would be very challenging and time-consuming. The solid state reactive sintering lead to the preparation of superconducting samples with good Tc onset but containing several impurities, which compromised the steepness of transition and the current carrying capability. This route requires further optimization, which can be achieved keeping cleaner all the process steps. Three routes were implemented for the preparation of FeSeTe samples, the solid state reactive synthesis, the mechano-chemical synthesis and the synthesis by fusion. The first two routes, as happened for FeSe samples, need further optimization. The third route brought to the preparation of several very good polycrystalline samples by a melting process, with heat treatment (HT) at temperatures of about 970 °C followed by cooldown to about 400 °C. It was verified that, as a consequence of the fusion process, impurities and spurious phases between grains are mostly removed, a preferential orientation of the samples is promoted and the critical current is enhanced. Therefore this fabrication route is recommended in view of applications, even if further efforts are needed to develop the material ready to use for example as a target for films deposition or eventually for the preparation of actual strands. In this work the main physical characterizations performed on all kinds of produced samples are shown. The reproducibility of the superconducting properties of samples prepared with the same procedure has been verified and only the representative samples for each group have been shown for clarity and readability. In particular the performing samples have been object of an extensive characterization, carried out in different superconducting labs at ENEA CR Frascati, at Master lab of CNR-SPIN Salerno and Physics Department of University of Salerno. Beside structural, magnetic, transport and calorimetric measurements, several analysis concerning the pinning mechanisms acting and competing inside the produced samples have been performed, within the framework of several literature models. As expected, pinning properties strongly depend on the preparation procedures which induces the defect structure into the samples. Magnetic relaxation measurements have supported this analysis, giving a corroborating possible interpretation of the measured peak effect, if present, and to the behaviour of the effective energy barrier as a function of the current density. In conclusion, despite the undeniable polycrystalline nature of the FeSeTe samples, those obtained by melting process present superconducting properties closely resembling the single crystals ones, with onset temperatures of about 15 K and quite steep transitions. Best performing samples have large hysteresis cycles well opened up to 12 T (at about 9 K) and up to 18 T (at about 7 K) with a robust critical current density weakly dependent on the applied field in the high field range. [edited by author]<br>Nella comunità scientifica c’è grande interesse per quanto riguarda l’esplorazione di nuovi materiali superconduttori adatti per applicazioni ad alto campo, con l’obiettivo di venire incontro alle incalzanti necessità dell’industria. In questo contesto, i materiali superconduttori a base di ferro, scoperti recentemente, rappresentano un’opzione promettente, soprattutto grazie alla loro temperatura critica, intermedia rispetto a quelle dei superconduttori a bassa e ad alta Tc, e anche grazie ai loro campi magnetici superiori estremamente alti. L’obiettivo di questo lavoro è stata la preparazione e lo studio delle proprietà fisiche di campioni superconduttori di a base di ferro-calcogenuri, in particolare di FeSe e FeSeTe policristallino. La famiglia dei ferro-calcogenuri è stata scelta soprattutto a causa delle sue interessanti proprietà superconduttive, e anche a causa della sua semplice struttura cristallina e alla mancanza di elementi tossici nella sua composizione. Ho portato avanti diverse strade per la produzione dei campioni, aprendo così un campo di ricerca completamente nuovo all’ENEA di Frascati. Ho conseguito parte delle competenze necessarie per questo lavoro lavorando anche presso altri laboratori che vantano grande esperienza nella preparazione di materiali superconduttori a base di ferro, in particolare presso i laboratori del National Institute of Material Science (NIMS) di Tsukuba in Giappone, dove ho lavorato presso il Nano Frontier Material Group, guidato dal Prof. Dr. Takano. Ho avuto anche l’opportunità di lavorare un breve periodo presso i laboratori del CNR SPIN di Genova e presso i dipartimenti di Fisica e Chimica dell’Università di Genova, dove ho potuto confrontarmi con ricercatori esperti nella produzione di campioni superconduttori a base di ferro. La maggior parte delle competenze le ho comunque acquisite per esperienza diretta. Anche se alcune delle strade intraprese per la preparazione dei campioni non hanno portato ai risultati sperati, alcune di queste tecniche hanno dato risultati interessanti, altre meritano senz’altro un ulteriore ottimizzazione. Per quanto riguarda i campioni in FeSe, sono stati implementati due processi per la preparazione: la deposizione elettrochimica su substrato di ferro e la sintesi allo stato solido. Il primo processo ha portato come risultato film sottili di FeSe contenenti la fase tetragonale superconduttiva, ma l’ottimizzazione delle proprietà superconduttive di questi campioni sarebbe un procedimento molto lungo e complesso. La sintesi allo stato solido ha portato alla preparazione di campioni superconduttivi con una buona Tc di onset ma contenenti molte impurezze, che hanno compromesso la ripidità della transizione e le proprietà di trasporto di corrente. Questa tecnica di preparazione richiede un’ottimizzazione futura, che potrà essere ottenuta mantenendo più pulite tutte le fasi del processo. Per la preparazione del FeSeTe sono state utilizzate tre tecniche, la sintesi allo stato solido, la sintesi meccano-chimica e la sintesi per fusione. Le prime due strade, come visto per i campioni in FeSe, hanno bisogno di essere ulteriormente ottimizzate. La terza strada ha portato alla preparazione di diversi campioni policristallini di qualità molto buona tramite un processo di fusione, con trattamenti termici a temperature di circa 970°C seguiti dl raffreddamenti a circa 400°C. È stato verificato che, come conseguenza del processo di fusione, le impurità e le fasi spurie tra i grani sono quasi totalmente rimosse, è promosso l’orientamento preferenziale dei campioni e la corrente critica risulta migliorata. Quindi questa tecnica per la preparazione dei campioni è quella raccomandata in vista delle applicazioni, anche se ulteriori sforzi saranno necessari per sviluppare un materiale pronto per l’uso ad esempio come target per la deposizione di film o eventualmente per la preparazione di veri e propri fili. In questo lavoro di tesi sono state mostrate le principali caratterizzazioni eseguite su tutti i tipi di campioni prodotti. La riproducibilità delle proprietà superconduttive dei campioni preparati con le stesse procedure è stata verificata e solo i campioni rappresentativi per ogni gruppo sono stati mostrati, per chiarezza e per non compromettere la leggibilità della tesi. In particolare i campioni migliori sono stati oggetto di una caratterizzazione estesa, portata avanti in vari laboratori di superconduttività, all’ENEA di Frascati, al Master Lab del CNR SPIN di Salerno e al dipartimento di Fisica dell’Università di Salerno. Oltre alle caratterizzazioni strutturali, magnetiche, di trasporto e calorimetriche, sono state portate avanti diverse analisi riguardanti i meccanismi di pinning che agiscono e competono all’interno dei campioni prodotti, nell’ambito di diversi modelli tratti dalla letteratura scientifica relativa all’argomento. Come previsto, le proprietà di pinning dipendono fortemente dalle procedure di preparazione che inducono e influenzano la struttura dei difetti all’interno dei campioni. Queste analisi sono state supportate anche da misure di rilassamento magnetico, che hanno corroborato l’interpretazione del peak effect, quando presente, e il comportamento dell’energia di attivazione di barriera in funzione della densità di corrente. In conclusione, nonostante l’innegabile natura policristallina dei campioni di FeSeTe preparati, quelli ottenuti per fusione presentano proprietà superconduttive che ricordano da vicino quelle dei campioni cristallini, con temperature di onset di circa 15 K e transizioni piuttosto strette. I campioni migliori hanno infatti cicli di isteresi ben aperti fino a 12 T (a circa 9 K) e fino a 18 K (a circa 7 K) con una densità di corrente critica che dipende debolmente dal campo magnetico nell’intervallo degli alti campi. [a cura dell'autore]<br>XIV n.s.

Magnetic reconnection is a fundamental process in highly conducting fluids and plasmas whereby the magnetic field line topology is rearranged and magnetic energy is converted into thermal energy, bulk kinetic energy and fast particle energy. It has been widely recognized to play an essential role in many events occurring in laboratory plasmas, classical examples of which are sawtooth crashes and disruptions in fusion devices, as well as in space and astrophysical plasmas, with magnetospheric substorms and solar flares being the most prominent examples. For this reason magnetic reconnection has attracted increasing consideration in recent years. Indeed, to deepen the knowledge of the microscopic scale reconnection physics, a Magnetospheric MultiScale (MMS) mission has been planned by NASA. Furthermore, several dedicated laboratory experiments have been designed in the last decade with the aim to advance the understanding of reconnection phenomena in regimes of interest for fusion plasmas. The modeling of magnetic reconnection in these regimes, characterized by low particle collisionality and high temperature, should require a kinetic description. On the other hand, important effects occurring at kinetic scales as the gyro-radii and particle inertial lengths can be described within a generalized fluid description, which is particularly desirable because of the physical intuition, the analytical tractability and the computational gain. For these reasons, in this thesis we have modeled magnetic reconnection adopting the latter approach. The analytical analysis and numerical simulations performed in this work have allowed us to obtain new results on the behaviour of magnetic reconnection in high-temperature plasmas. In particular, we have found new dispersion relations for the growth rate of the reconnecting instability in the presence of an equilibrium flow velocity, and we have also shown the relevance of the ion gyration on the growth rate and the field structures characterizing fast reconnection phenomena. The most remarkable result consists in having found that the gyro-motion of hot ions causes a novel acceleration phase of the reconnection process, which may help with the interpretation of experimental observations.

The always more predominant synergy between the biological and engineering worlds, is leading to an extremely fruitful mutual cooperation in which biology can benefit of engineering innovations and, vice versa. In this framework, unknown biological processes and complex bio-chemical dynamics can finally be unveiled, tremendously contributing to improve the current health-care and diagnostic paradigm. To reach this ambitious aim, it is fundamental to understand how these two worlds, apparently so distant, can efficiently and mutually integrate. My Ph.D. thesis was focused, specifically, on the interaction between electronics and cardiac cells, due to the intrinsic electrical activity of the latter. Due to the micrometer dimensional scale of the cells, it is advantageous to also use micrometer-scale electrodes, capable to interface with the living entities at their natural and functional dimensions where the most relevant processes happen. For this reason, I analyzed in depth also the use of electrical platforms, embedding micro-electrodes, both passive or with CMOS circuits directly underneath. Due to the complex nature of the problem of the cells viability on-chip, at first synthetic microstructures were used to test and optimize the sensing capabilities of the developed platforms. Specifically, zinc oxide (ZnO) micro-wires were at first deposited across the nanometric-spaced gap induced by exploiting the electromigration phenomenon on a micrometric gold metal wire. Thanks to the physical/electrical properties of ZnO microwires, the main goal was to create a multi-sensor capable to simultaneously detect variations of pH, temperature and UV-visible light, all important parameters to monitor in cell cultures. To improve the robustness of the system against electrical noise, the same microstructures were deposited across couples of micrometer electrodes, but this time fabricated on the surface of a CMOS chip. In this way, the electronic read-out of the variation of the electrical properties of the microstructure, according to the variation of the external parameters selected, was performed directly underneath the electrodes surface, hence reducing noise and parasitics. Specifically, 24 couples of aluminum electrodes, were designed so to allow multiple measurements on the same chip. To improve the quality of the electrical contact between the microstructure and the aluminum electrode, easily oxidizing, a tailored electroless gold plating process was optimized to modify the electrodes surface to gold. The CMOS chip with deposited ZnO microwires was then used to investigate the UV-visible light sensing capabilities of the microstructure. To understand the complex and multiple requirements of performing, instead, measurements of living cardiac cells on chip, a visiting period of 6 months was spent at the Kademhosseini laboratory in Boston (USA), part of the Harvard-MIT health science and technology division. During this time, the needs of primary neonatal rats cardiomyocytes were investigated by developing and electrically modulated bio-hybrid actuator, triggered by the beating activity of cardiac cells, and fabricated using bio-compatible hydrogels. After this extremely important intermediate step, preliminary experiments were done, at the end of the Ph.D. activity by measuring the electrical signals of cardiac cells directly on the surface of CMOS multi-electrodes arrays platforms. Some tests were performed also functionalizing the surface of the CMOS chips with the hydrogel developed in Boston for the bio-hybrid actuator. This activity was performed at the Italian Institute of Technology of Genova (Italy). The main focus of this activity was trying to evoke the firing of an electrical cardiac signal, commonly known as cardiac action potential, by applying an electrical stimulation. At the same time, the biologically-produced electrical signals were also recorded in real-time. Although the performed activity represented only an initial proof-of-concept, it provided an important insight to the challenges to be faced when living entities are deposited on electronic chips. At the same time, the preliminary experiments revealed the huge potentiality of active micro-electronic platforms for the monitoring, study and excitation of living cardiac cells. Thanks to these extremely multi-disciplinary studies it will be possible to significantly impact the current health-care and drug-development paradigms due to the possibility to investigate, previously unknown or not well characterized, biological processes, directly at their proper dimensional scale, with state-of-the-art integrated and hybrid electronic platforms.