Time-lapse fluorescence live cell imaging has been widely used to study various dynamic processes in cell biology. As the initial step of image analysis, it is important to localize and segment cell edges with higher accuracy. However, fluorescence live-cell images usually have issues such as low contrast, noises, uneven illumination in comparison to immunofluorescence images. Deep convolutional neural networks, which learn features directly from training images, have successfully been applied in natural image analysis problems. However, the limited amount of training samples prevents their routine application in fluorescence live-cell image analysis. In this thesis, by exploiting the temporal coherence in time-lapse movies together with VGG-16 [1] pre-trained model, we demonstrate that we can train a deep neural network using a limited number of image frames to segment the entire time-lapse movies. We propose a novel framework, vU-net, which combines the advantages of VGG-16 [1] in feature extraction and U-net [2] in feature reconstruction. Moreover, we design an auxiliary convolutional block at the end of the architecture to enhance edge detection. We evaluate our framework using dice coefficient and the distance between the predicted edge and the ground truth on high-resolution image datasets of an adhesion marker, paxillin, acquired by a Total Internal Reflection Fluorescence (TIRF) microscope. Our results demonstrate that, on difficult datasets: (i) The testing dice coefficient of vU-net is 3.2% higher than U-net with the same amount of training images. (ii) vU-net can achieve the best prediction results of U-net with one third of training images needed by U-net. (iii) vU-net produces more robust prediction than U-net. Therefore, vU-net can be more practically applied to challenging live cell movies than U-net since it requires a small size of training sets and achieved accurate segmentation.

This diploma thesis deals with design, implementation and testing of fluorescence microscope intended for carbon nanotubes spectroscopy. Theory of fluorescence resulting from solid state physics and atomic structure of nanotubes is briefly discussed. The basic idea, optical scheme and optical elements used in the fluorescence microscope are described based on requirements resulting from the theory. The thesis mentions the sample preparation procedure and measurements of its optical activity as well. Realized microscope is usable for fluorescence measurements within the supposed range of wavelengths as was proved by testing.

The aim of this diploma work is to improve selectivity and sensitivity in DNA-chips by utilizing fluorescence resonance energy transfer (FRET) between conjugated polyelectrolytes (CPEs) and fluorophores. Leclerc and co-workers have presented successful results from studies of super FRET between fluorophore tagged DNA and a CPE during hybridisation of the double strand. Orwar and co-workers have constructed a DNA-chip using standard photo lithography creating a pattern of the hydrophobic photoresist SU-8 and cholesterol tagged DNA (chol-DNA). This diploma work will combine and modify these two ideas to fabricate a improved DNA-chip. Immobilizing of DNA onto surface has been done by using soft lithography. Hydrophobic pattern arises from the poly(dimethylsiloxane) (PDMS) stamp. The hydrophobic pattern will attract chol-DNA that is adsorbed to the chip. Different sets of fluorophores are covalently bound to the DNA and adding CPEs to the complex will make FRET occur between CPE and bound fluorophore. We will here show that the specificity in DNA hybridization by using PDMS patterning was high. FRET clearly occurred, especially with the CPEs as donor to the fluorophore Cy5. The intensity of FRET was higher when the fluorophore and the CPE were conjugated to the same DNA strand. The largest difference in FRET intensity between double stranded and single stranded complexes was observed with the CPE tPOMT. Super FRET has been observed but not yet fully proved. The FRET efficiency was lower with the fluorophore Alexa350 as donor compared to the Cy5/CPE complex. Most of the energy transferred from Alexa350 was extinguished by quenching.

This research presents a detection scheme for analyzing the temperature distribution produced by the Joule heating effect nearby the channel wall in a microfluidic chip utilizing a temperature-dependent fluorescence dye. An advanced optical microscope system¡Xtotal internal reflection fluorescence microscope (TIRFM) is used for measuring the temperature distribution on the inner channel wall at the point of electroosmotic flow in an electrokinetically driven microfluidic chip. In order to meet the short working distance of the objective-type TIRFM, microscope cover glass are used to fabricate the microfluidic chips. The short fluorescence excitation depth from a TIRFM makes the intensity information obtained is not sensitive to the channel depth variation which ususally biases the measured results while using conventional epi-fluorescence microscope (Epi-FM). Therefore, a TIRFM can precisely describe the temperature profile of the distance within hundreds of nanometer of the channel wall where consists of the Stern layer and the diffusion layer for an electrokinetic microfluidic system. In order to investigate the temperature distribution produced by the Joule heating effect for electrokinetically driven microchips, this study not only measures the temperature on the microchannel wall by the proposed TIRFM but also measures the temperature inside the microchannel by an Epi-FM. In addition, this research presents a method to reduce the Joule heating effect and enhance the separation efficiency of DNA biosamples in a chip-based capillary electrophoresis (CE) system utilizing pulse DC electric fields. Since the average power consumption is reduced by the pulse electric fields, the Joule heating effect can be significantly reduced. Results indicate the proposed TIRFM method provides higher measurement sensitivity over the Epi-FM method. Significant temperature difference along the channel depth measured by TIRFM and Epi-FM is experimentally observed. In addition, the measured wall temperature distributions can be the boundary conditions for numerical investigation into the Joule heating effect. The proposed method gives a precise temperature profile of microfluidic channels and shows the substantial impact on developing a simulation model for precisely predicting the Joule heating effect in microfluidic chips. Moreover, in the research of reducing the Joule heating effect and enhancing the separation efficiency in a chip-based CE system utilizing pulse electric fields, the experimental and numerical investigations commence by separating a mixed sample comprising two fluoresceins with virtually identical physical properties. The separation level is approximately 2.1 times higher than that achieved using a conventional DC electric field. The performance of the proposed method is further evaluated by separating a DNA sample of Hae III digested £XX¡V174 ladder. Results indicate the separation level of the two neighboring peaks of 5a (271 bp) and 5b (281 bp) in the DNA ladder is as high as 120% which is difficult to be achieved using a conventional CE scheme. The improved separation performance is attributed to a lower Joule heating effect as a result of a lower average power input and the opportunity for heat dissipation during the zero-voltage stage of the pulse cycle. Overall, the results demonstrate a simple and low-cost technique for achieving a high separation performance in CE microchips.

<p>The aim of this diploma work is to improve selectivity and sensitivity in DNA-chips by utilizing fluorescence resonance energy transfer (FRET) between conjugated polyelectrolytes (CPEs) and fluorophores.</p><p>Leclerc and co-workers have presented successful results from studies of super FRET between fluorophore tagged DNA and a CPE during hybridisation of the double strand. Orwar and co-workers have constructed a DNA-chip using standard photo lithography creating a pattern of the hydrophobic photoresist SU-8 and cholesterol tagged DNA (chol-DNA). This diploma work will combine and modify these two ideas to fabricate a improved DNA-chip.</p><p>Immobilizing of DNA onto surface has been done by using soft lithography. Hydrophobic pattern arises from the poly(dimethylsiloxane) (PDMS) stamp. The hydrophobic pattern will attract chol-DNA that is adsorbed to the chip. Different sets of fluorophores are covalently bound to the DNA and adding CPEs to the complex will make FRET occur between CPE and bound fluorophore.</p><p>We will here show that the specificity in DNA hybridization by using PDMS patterning was high. FRET clearly occurred, especially with the CPEs as donor to the fluorophore Cy5. The intensity of FRET was higher when the fluorophore and the CPE were conjugated to the same DNA strand. The largest difference in FRET intensity between double stranded and single stranded complexes was observed with the CPE tPOMT. Super FRET has been observed but not yet fully proved. The FRET efficiency was lower with the fluorophore Alexa350 as donor compared to the Cy5/CPE complex. Most of the energy transferred from Alexa350 was extinguished by quenching.</p>

To supplement and modify the diagnosis and clinical research of B-cell Chronic Lymphocytic Leukemia (B-CLL), a new method based on cell imaging and image processing was developed and applied to the B-CLL patient samples. The fluorophore-labelled leukemia cells were clearly visualized, reflecting the positive/negative expression of the corresponding surface markers and their distribution. Computer algorithms were devised and used to analyze a large number of images. The fluorescence intensity of the labelled antibodies on a given cell directly reflects the expression of the corresponding surface markers. The morphology and size of leukemia cells were not identical even in the same patient’s sample and the size variation does not correlate with the number of surface markers. The amount of each surface marker was approximately fixed for each patient, but there were some relationships, for instance, the number of CD19 and CD38 markers were correlated to each other. The heterogeneous expression of surface markers confirmed an assumption that surface markers have their preferred membrane positions. One of the most important results is that the cell imaging and our image processing method has provided an alternative and reliable way to diagnose B-CLL and new insights in the prognosis of subtype of B-CLL.

To supplement and modify the diagnosis and clinical research of B-cell Chronic Lymphocytic Leukemia (B-CLL), a new method based on cell imaging and image processing was developed and applied to the B-CLL patient samples. The fluorophore-labelled leukemia cells were clearly visualized, reflecting the positive/negative expression of the corresponding surface markers and their distribution. Computer algorithms were devised and used to analyze a large number of images. The fluorescence intensity of the labelled antibodies on a given cell directly reflects the expression of the corresponding surface markers. The morphology and size of leukemia cells were not identical even in the same patient’s sample and the size variation does not correlate with the number of surface markers. The amount of each surface marker was approximately fixed for each patient, but there were some relationships, for instance, the number of CD19 and CD38 markers were correlated to each other. The heterogeneous expression of surface markers confirmed an assumption that surface markers have their preferred membrane positions. One of the most important results is that the cell imaging and our image processing method has provided an alternative and reliable way to diagnose B-CLL and new insights in the prognosis of subtype of B-CLL.

This diploma thesis deals with the possibilities of stress resistance analysis of the Cupriavidus necator H16 and PHB-4 bacterial cells by spectroscopic methods and by testing the suitability of acridine orange as a viable dye. Based on research in literature, suitable analytical methods have been proposed, namely flow cytometer and fluorescence microscope. The first part of the experimental work was focused on the fluorescence microscope, which confirmed the basic character of acridine orange. Three stress factors, 50% and 70% ethanol, and acidic pH (pH = 1) were selected for viability monitoring. The bacteria fluoresced with green color after exposure to ethanol and red spots were found next to the cells, indicating their loss of integrity. In an acidic environment, the bacteria fluoresced red because of a partial DNA breakdown. The results were verified by the combination of propidium iodide with SYTO9 and the acridine orange suitability proved to be useful in this method. Image records were processed using image analysis. In the second part, acridine orange was used to monitor fluorescence using a flow cytometer. The result of the measurement was fluorescence expressed as histograms for individual channels, where fluorescence was characterized by median and mean intensity. By comparing the methods used, the acridine orange appears to be a more suitable fluorescent dye for the microscope than for a flow cytometer in which it was more difficult to obtain cell viability information. In the last part of the experimental work interesting photophysical properties of acridine orange were investigated.

A relatively new application for the laser is in fluorescence microscopes. The fluo- rescence microscope needs a high power light source input. Using a laser source improves the precision of the microscope. A pulsed laser source enhances the performance of the fluorescence microscope and a laser diode can be overdriven without being damaged. The thesis investigates which properties of the laser pulses are needed regarding pulse width, pulse period and waveform. The thesis also investigates which properties are desired for the electrical pulses driving the laser, and how they can be generated using electrical components. The desired laser pulse should have a pulse width of 100 ps and a pulse period of 50 ns. The laser pulse should also have a well-defined wavelength, stable output power and it should be able to quickly turn on and off. To achieve this laser pulse, the desired input to the laser diode should have an input voltage of 5 V, an input current of 250 mA, a pulse width of 100 ps and a pulse period of 50 ns. For generating this pulse the chosen pulse generator, an SRD, should have low junction capacitance, low package capacitance and low package inductance. The chosen amplifier, a MESFET, desires low drain current and should have high transconductance and a large negative threshold voltage.

No description available.