Abstract (eng)
Peroxisomes are single membrane bound organelles enclosing various important metabolic pathways. The peroxisomal matrix protein import is mediated by two targeting signals namely peroxisomal targeting signal 1 (PTS1, C-terminal) and peroxisomal targeting signal 2 (PTS2, N-terminal). The PTS1 is bound by the receptor PEX5 and PTS2 is bound by a dimer consisting of the receptor PEX7 and co-receptor PEX5 which are then able to cross the peroxisomal membrane.
The efficiency of peroxisomal matrix protein import depends on the quality of the PTS. To determine the quality of individual PTS, a visual evaluation of import efficiency in cells expressing reporter proteins has originally been used. In this study, we want to present a first step into the direction of automated image analysis of fluorescence images. On the one hand, peroxisomes were first identified by a segmentation algorithm called Trainable Weka Segmentation and the average of the intensity maxima of these areas was compared to the average intensities of the non-peroxisomal cellular background. On the other hand, the analysis of the intensity distribution of all pixels of each cell was used to discriminate cells presenting with different import efficiencies of the reporter protein, independently of the different shapes of cells. The results of the segmentation approach demonstrates significant differences between the mean intensities of peroxisomal and non-peroxisomal areas within cells. The results of the analysis of the pixel-based intensity distribution suggest that the shape of the distribution curves reflect the visual evaluation of import efficiency. Moreover, the total amount of the reporter protein has a great impact on the evaluation of import efficiency. In summary, both methods presented in this thesis show promising results and are good candidates for further testing and improvement.