Knowing the localisation and spatial organisation of proteins is crucial for understanding their function. The development of super-resolution imaging has improved our ability to garner this information, but counting individual molecules in densely-packed assemblies is still challenging. DNA-based point accumulation for imaging in nanoscale topography (DNA-PAINT) is one of the most recently developed imaging techniques in super-resolution microscopy. It uses fluorescently-labelled DNA to visualise the molecules of interest with nanometre precision. DNA-PAINT was initially reliant on antibody labelling of in vitro protein targets, however, there is need for an alternative labelling strategy as good antibodies do not exist for many target proteins. Moreover, it is impossible to quantify antibody labelling efficiency, which is a crucial parameter for quantitative imaging. In order to address these issues, I present here an optimised imaging pipeline for protein counting in a thick tissue sample, tens of microns away from the coverslip, for which cell polarity proteins in epithelial cells of the fruit fly (Drosophila melanogaster) egg chambers are given as an example. Firstly, I established an alternative labelling strategy to label polarity proteins for DNA-PAINT imaging using genetically-encoded Halo and SNAP self-labelling enzymes in fruit fly tissue. In this approach, the Halo and SNAP ligands conjugated to DNA react with their respective enzymes to form a covalent bond with the protein of interest in a 1:1 stoichiometry. I then optimised the labelling protocol for imaging the fixed fruit fly tissue and analysed non-specific signal to reduce background during image post-processing. A quantitative Western blot-based gel band shift assay was developed to determine the labelling efficiency of target proteins. Moreover, I used nucleoporin proteins in the nuclear pore complex to calibrate the influx rate of fluorescently-labelled DNA to quantify the number of molecules in super-resolution images. Additionally, I used nucleoporin-160 and nucleoporin-188 to benchmark two-colour super-resolution imaging using DNA-PAINT. Super-resolution imaging of three apical polarity proteins (aPKC, Crumbs, Par6) in the fruit fly egg chambers revealed that they form mesoscopic-sized clusters along the cell junctions. In order to analyse these clusters in a quantitative manner, I collaborated with Leila Muresan to develop an image analysis pipeline. My analysis demonstrated that apical polarity proteins are less concentrated in the cytosol by approximately one order of magnitude. To expand on these observations, the junctional clusters were identified by a mean-shift algorithm and classified according to size, i.e. the number of molecules. The cluster size distribution was then approximated by a mathematical function. The model selection was performed by Bayesian information criteria that was tested on simulated data beforehand. This work provides an optimised imaging pipeline for quantifying the number of protein molecules in a thick biological sample using DNA-PAINT, and proposes a post-processing approach to identify and mathematically describe molecular clustering. These data will prove useful for modelling the spatial organisation of polarity proteins, and provide a framework for greater insight into the biological function of individual proteins.