Figure 2

Image analysis framework for understanding pore-free islands. Our framework was constructed using the following three categories: image preparation, image processing (IP), and quantification. Image preparation shows the acquisition of three sets of co-immunostained images that were collected at the bottom-surface of NE. After manually specifying the nuclear regions in the mAb414-images using our image processing software (http://logistics.riken.jp/vcat/vcat/en), IP1 was next performed to recognise the area of pore-free islands in each specified nuclear region. The recognition results were subjected to statistical analysis of the pore-free island size with respect to inhibitor type, the time of inhibitor treatment, and primary antibodies (Quantification 1). Subsequently, the recognised areas of pore-free islands were used as the initial regions for automatic detection of Nups foci in the NUP93, NUP107, and POM121 images. Automatic Nups foci detection was performed using the Otsu method, focusing on the geometric feature (see Materials and Methods). According to the measurements of the detected Nups foci in each pore-free island, quantified data regarding their appearance frequency were obtained (Quantification 2). In addition, we investigated the position of Nups foci in the pore-free islands to determine whether their distribution was spatially biased by comparing the observed and simulated data (Quantification 3).