Modern commercial fluorescence microscopes allow to study cellular features on the nanoscopic level, using innovative optical principles and continuously improving sample preparation technologies. However, the complexity of an instrument and the associated price scale with the attainable resolution and the image quality. Open-source concepts such as the UC2 (You. See. Too.) microscopy toolbox have proven to be low-cost versatile alternative to adapt a microscope to the requirements of an experiment, allowing to interface various optical components which were initially exclusive to commercial systems. Unfortunately, the versatility and the use of low-cost components often results in a loss of image quality and resolution. To overcome this challenge, I incorporated a high quality objective into an UC2-based microscope assembled from easily available optics, affordable electronics and 3D printed parts. The setup is considerably cheaper than its commercial counterparts while providing a maximum of capabilities, thereby providing a proof of concept that high-quality quantitative data can be generated with a low-budget system. The modalities applied include high quality imaging of fxed and live cells in both fluorescent and bright-feld channels, as well as continuous measurements of live-cells inside an incubator, single particle tracking of membrane molecules and super-resolution imaging. In one application, single particle tracking of GPI anchored GFP in the cellular plasma membrane was performed and a diffusion coeffcient matching the values found in literature was determined. In a second application, single molecule localization microscopy was used to surpass the resolution limit imposed by light diffraction. Hereby, the extrapolated diameter of microtubules was in accordance with values measured on a higher cost setup. Low budget optics i.e. objectives have reduced light collection capability, are prone to optical aberrations and cannot correct irregularities within the wavefront. In an effort to further increase the setup’s accessibility, low-budget objective lenses were assessed to determine for which imaging assays they were employable. In a second part of the project, various tools have been used in order to facilitate the analysis of microscopy data, with emphasis on single molecule localization microscopy. Machine learning based techniques were tested to ensure the automation of image analysis pipelines, thereby aiming to reduce the demand for manual processing of large data-sets. Clustering of the single molecules positions is a technique that helps to understand the underlying pathways when studying the emergence of signaling events at the plasma membrane. Bayesian Cluster Analysis and visualization application is a software that facilitates the access to clustering through an user-friendly GUI, a reduced complexity for viithe user and an increased effciency of the computational load. After benchmarking the algorithm on simulations of clustered cell membranes, the clustering behaviour of the CD95 receptor was characterized on actual samples. Further experiments will investigate the receptor organisation upon ligand activation to gain insights into the pathways initiated by the receptor.
