Modern techniques like automated microscopy and spectroscopy now make it possible to study quantitatively, across multiple phenotypic and molecular parameters, how cell populations are affected by different treatments and/or environmental disturbances. As the technology development at the instrument level often is ahead of the data analytical tools and the scientific questions, there is a large and growing need for computational algorithms enabling desired data analysis. These algorithms must have capacity to extract and process quantitative dynamic information about how the cell population is affected by different stimuli with the final goal to transform this information into development of new powerful therapeutic strategies. In particular, there is a great need for automated systems that can facilitate the analysis of massive data streams for label-free methods such as phase contrast microscopy (PCM) imaging and spectroscopy (NMR). Therefore, in this thesis, algorithms for quantitative high-throughput phenotypic and systemic profiling of in vitro growing cell populations via label-free microscopy and spectroscopy are developed and evaluated. First a two-dimensional filter approach for high-throughput screening for drugs inducing autophagy and apoptosis from phase contrast time-lapse microscopy images is studied. Then new methods and applications are presented for label-free extraction and comparison of time-evolving morphological features in phase-contrast time-lapse microscopy images recorded from in vitro growing cell populations. Finally, the use of dynamic morphology and NMR/MS spectra for implementation of a reference database of drug induced changes, analogous to the outstanding mRNA gene expression based Connectivity Map database, is explored. In conclusion, relatively simple computational methods are useful for extraction of very valuable biological and pharmacological information from time-lapse microscopy images and NMR spectroscopy data offering great potential for biomedical applications in general and cancer pharmacology in particular.

No description available.

This work is aimed at deposition of fibers, which are able to penetrate deeper in the human respiratory airways and cause health hazards. Opening chapters are dedicated to classification of aerosols, human lung anatomy basics and methods for measurement of micrometer-sized aerosols. An experiment was carried out to investigate the deposition of fibers. Fibers were delivered into the silicon cast of the human airways and data of deposition fraction and efficiency were acquired. A new method was established to acquire the data. This method works on a principle of image analysis. Results of the new method were compared with a standard method, which follows a methodology NIOSH 7400.

Ce manuscrit détaille le développement d'un montage de pince optique permettant d'étudier les propriétés mécaniques des cellules endothéliales, impliquées dans le développement de l'athérosclérose. Le but est de déterminer les propriétés viscoélastiques des cellules, et de suivre la propagation d’une contrainte mécanique au sein de la cellule. Cette contrainte mécanique est appliquée via une bille liée à la membrane de la cellule et soumise à un piège optique.Le dispositif réalisé combine le piégeage optique et la microscopie à contraste de phase, permettant d'exercer une force tout en imageant les cellules via le même objectif de microscope. L'originalité du montage de pince optique repose sur la détection du signal rétrodiffusé par la bille piégée, dans un plan conjugué du plan focal arrière de l'objectif, afin de mesurer la position relative de la bille par rapport au piège.Une part importante de ce travail a consisté à comprendre l'allure du signal détecté présentant un système d'interférences en anneaux, et à l’expliquer par un modèle simple. Ce modèle a permis de comprendre la présence d’artefacts de mesure de position dus à la superposition de l'anneau de phase sur la figure d’interférence. Pour y remédier, l'anneau de phase est déporté dans un plan conjugué intervenant uniquement dans l'imagerie de l'échantillon.La figure d'interférence présente un atout majeur : elle donne accès à la hauteur précise de la bille piégée, généralement difficile à mesurer. Cette information est nécessaire pour calibrer la constante de raideur du piège optique à la hauteur des cellules, que ce soit par l'analyse de la densité spectrale de puissance du mouvement brownien de la bille piégée ou par sa réponse à un échelon de position du piège. Ces deux méthodes de calibration, ainsi que l'application du théorème d’équipartition et l'analyse par inférence bayésienne, ont été mises en œuvre. Tous les résultats s'avèrent en bon accord. La calibration complète du dispositif en fait un outil prêt à l'emploi pour exercer des forces locales contrôlées en direction et en amplitude sur les cellules. / This manuscript details the development of an optical tweezer setup to study the mechanical properties of endothelial cells, involved in the development of atherosclerosis. The goal is to determine the viscoelastic properties of the cells, and to follow the propagation of the mechanical constraint inside the cell. This mechanical constraint is applied via a bead attached to the cell membrane and subjected to an optical trap.The setup built combines optical trapping with phase contrast microscopy, to apply a force while imaging the cells with the same microscope objective. The originality of the optical tweezer setup relies on the detection of the signal backscattered by the trapped bead, in a plane conjugate to the back focal plane of the objective, in order to measure the relative position of the bead with respect to the center of the trap.An important part of this work was dedicated to the understanding of the detected signal presenting an interference pattern with rings, explained by a simple model. This model provides an explanation for the position measurement artifacts arising from the superposition of the phase ring and the interference pattern. To solve the problem, the phase ring was moved in a conjugate plane involved only in the imaging path of the sample.The interference pattern has the main advantage of giving access to the precise height of the trapped bead, usually difficult to measure. This information is necessary to calibrate the optical trap stiffness at the height of the cells, either by the power spectrum analysis of the Brownian motion of the trapped bead, or by its response to a step motion of the trap. These two calibration methods, along with the application of the equipartition theorem and Bayesian inference analysis, were implemented and their results compared, showing a good agreement. The complete calibration of the setup makes it a ready-to-use tool to exert local forces controlled in direction and amplitude on cells.

This dissertation deals with customised image analysis tools in cancer research. In the field of biomedical sciences, mathematical imaging has become crucial in order to account for advancements in technical equipment and data storage by sound mathematical methods that can process and analyse imaging data in an automated way. This thesis contributes to the development of such mathematically sound imaging models in four ways: (i) automated cell segmentation and tracking. In cancer drug development, time-lapse light microscopy experiments are conducted for performance validation. The aim is to monitor behaviour of cells in cultures that have previously been treated with chemotherapy drugs, since atypical duration and outcome of mitosis, the process of cell division, can be an indicator of successfully working drugs. As an imaging modality we focus on phase contrast microscopy, hence avoiding phototoxicity and influence on cell behaviour. As a drawback, the common halo- and shade-off effect impede image analysis. We present a novel workflow uniting both automated mitotic cell detection with the Hough transform and subsequent cell tracking by a tailor-made level-set method in order to obtain statistics on length of mitosis and cell fates. The proposed image analysis pipeline is deployed in a MATLAB software package called MitosisAnalyser. For the detection of mitotic cells we use the circular Hough transform. This concept is investigated further in the framework of image regularisation in the general context of imaging inverse problems, in which circular objects should be enhanced, (ii) exploiting sparsity of first-order derivatives in combination with the linear circular Hough transform operation. Furthermore, (iii) we present a new unified higher-order derivative-type regularisation functional enforcing sparsity of a vector field related to an image to be reconstructed using curl, divergence and shear operators. The model is able to interpolate between well-known regularisers such as total generalised variation and infimal convolution total variation. Finally, (iv) we demonstrate how we can learn sparsity promoting parametrised regularisers via quotient minimisation, which can be motivated by generalised Eigenproblems. Learning approaches have recently become very popular in the field of inverse problems. However, the majority aims at fitting models to favourable training data, whereas we incorporate knowledge about both fit and misfit data. We present results resembling behaviour of well-established derivative-based sparse regularisers, introduce novel families of non-derivative-based regularisers and extend this framework to classification problems.

Semiconductors are used in many electronic applications, for example diodes, solar cells and transistors. Typically, semiconductors are inorganic materials, such as silicon and gallium arsenide, but lately more research and development has been devoted to organic semiconductors, for example semiconducting polymers. One of the reasons is that polymers can be customized, to a greater extent than inorganic semiconductors, to create a material with desired properties. Often, two polymers are blended to obtain the desired function, but two polymers do not usually result in an even blend. Instead they tend to separate from each other to varying degrees. The morphology of the blend affects the material properties, for example how efficiently it can convert electricity to light. In this project, thin films consisting of polymer blends were examined using microscopy techniques for the purpose of increasing our understanding of the morphology of such blends. One goal was to investigate whether a technique called correlative light and electron microscopy can be useful for examining the morphology of these films. In correlative light and electron microscopy, a light microscope and an electron microscope are used in the same location in order to be able to correlate the information from the two microscopes. The second goal was to learn about the morphology of the thin films using various microscopy techniques. The polymers used were Super Yellow and poly(ethylene oxide) with large molecular weight. Super Yellow is a semiconducting and light-emitting polymer while poly(ethylene oxide) is an isolating and non-emitting polymer that can crystallize. In the blend films, large, seemingly crystalline structures appeared. The structures could be up to 1 mm in the lateral direction, while the films were only approximately 170 nm thick. These structures could grow after the films had dried and their shapes were similar to those of poly(ethylene oxide) crystals. Consequently, there is reason to believe that it is the poly(ethylene oxide) that makes up the seemingly crystalline structures, but the structures also emitted more light than the rest of the film, and Raman spectroscopy showed that there was Super Yellow in the same location as the crystals. Among the microscopy techniques used, phase contrast microscopy was particularly interesting. This method visualizes differences in optical path length and was useful for studying polymer blends when the polymers have different indices of refraction. Correlating light and electron microscopy showed that there was a pronounced topographical difference between the seemingly crystalline regions and the rest of the thin film. Light microscopy has a limited resolution due to diffraction, but as long as the resolution of the light microscope is sufficient for seeing phase separation, correlative light and electron microscopy turned out to be a good method for studying the morphology of thin films of polymer blends. / Halvledare är viktiga för många elektroniska ändamål eftersom de kan användas till exempelvis dioder, solceller och transistorer. Traditionellt används inorganiska halvledande material som kisel eller galliumarsenid, men på senare tid har allt mer forskning och utveckling inriktat sig mot organiska (kolbaserade) halvledare, såsom halvledande polymerer, bland annat eftersom det i högre utsträckning går att skräddarsy de organiska materialen så att de får önskvärda egenskaper. Ofta blandas två polymerer med varandra för att skapa ett material med nya egenskaper som är önskvärda, men två polymerer brukar inte blandas jämnt utan tenderar att separera från varandra i olika utsträckning. Hur blandningen ser ut (morfologin) påverkar materialets egenskaper, till exempel hur effektivt det omvandlar ström till ljus. Med syfte att öka förståelsen för hur morfologin ser ut hos en blandning av två polymerer, har detta projekt gått ut på att undersöka tunna filmer av polymer-blandningar med hjälp av mikroskopiska tekniker. Ett delmål var att ta reda på om en teknik som heter korrelativ ljus- och elektronmikroskopi är en bra metod för att undersöka morfologin hos dessa filmer. Vid korrelativ ljus- och elektronmikroskopi används både ett ljusmikroskop och ett elektronmikroskop på samma plats för att kunna korrelera informationen som de båda mikroskopen ger. Det andra delmålet var att undersöka vad de olika mikroskopi-teknikerna kan säga om morfologin hos de tunna filmerna. De polymerer som använts är Super Yellow och poly(etylenoxid) med hög molekylmassa. Super Yellow är en oordnad halvledande och ljusemitterande polymer medan poly(etylenoxid) är en isolerande och icke-emitterande polymer som kan kristallisera. I de blandade filmerna uppstod stora kristall-liknande strukturer som kunde vara upp emot 1 mm breda trots att filmerna bara var ungefär 170 nm tunna. Dessa strukturer kunde växa fram efter det att filmerna redan hade torkat och påminde i form om kristaller som kan bildas av poly(etylenoxid). Det finns alltså skäl att tro att det är poly(etylenoxid) som kristalliserats, men de kristall-liknande strukturerna visade sig emittera mer ljus än vad resten av filmen gjorde, och Raman-spektroskopi visade att det även fanns Super Yellow på samma plats som kristallerna. Bland de mikroskopitekniker som testades utmärker sig faskontrastmikroskopi, som visar skillnader i den optiska vägskillnaden (det vill säga faktisk vägskillnad multiplicerat med brytningsindex). Det visade sig vara en intressant teknik för att studera polymerblandningar när de båda polymererna har olika brytningsindex. Genom att korrelera ljus- och elektronmikroskopi visade det sig att det fanns en tydlig skillnad i struktur mellan de kristall-liknande områdena och resten av den tunna filmen. Ljusmikroskopi har begränsad upplösning på grund av ett fenomen som heter diffraktion, men så länge som ljusmikroskopets upplösning är tillräcklig för att se fasseparation visade det sig att korrelativ ljus- och elektronmikroskopi är en bra metod för att studera morfologin hos tunna filmer av polymerblandningar.