Segmenting and analyzing cellular morphologies of Caenorhabditis elegans embryo with modular CShaperApp

Since C. elegans was adopted as a model organism, it has become a popular subject for numerous studies in developmental biology. From the fertilized egg to the end of embryogenesis, with approximately 550 living cells, each cell of C. elegans can be unambiguously named and tracked. Its embryogenesis can be visually inspected in vivo using advanced fluorescence microscopy, which captures different cellular structures like the cell nucleus and cell membrane. However, the large-scale imaging data with a relatively low signal-to-noise ratio imposes serious challenges in systematic documentation and investigation of the developmental process. Although several previous works proposed customized algorithms to segment C. elegans embryo images at the single-cell level, user-friendly software for morphological analysis is still urgently needed.

To fill the gap mentioned above, a collaborative team from Yan Lab at City University of Hong Kong, Tang Lab from Peking University and Zhao Lab at Hong Kong Baptist University, proposed an automatic cellular segmentation framework CShaperApp, to extract cellular morphological features up to the 350-cell stage in C. elegans embryogenesis. To improve the usability of previous work CShaper to this community, they developed desktop software CShaperApp that interactively performs image segmentation and morphological analysis. Modular functions of CShaperApp allow researchers to either adaptively retrain CShaper with their own dataset or process external datasets with our pretrained model.

Figure 1 illustrates the five modules of CShaperApp, inclusing Preprocess, Segmentation, Analysis, Result, and Train. The Preprocess module stacks all microscopic slices at each time point into one volumetric image (Figure 1A). When the cell position and lineage file, generated from StarryNite and AceTree, is available, the cell (nucleus) position information is extracted to synthesize binary nucleus masks where foreground pixels correspond to the centers of the cell nucleus. The Segmentation module adopts the CShaper framework to segment fluorescence images of cell membranes, transforming them into digitized segmentations with clear identification of the region corresponding to each cell (Figure 1B). The Analysis module automatically collects cellular morphological features such as cell volume, cell surface area, and cell-cell contact area, in the format of Excel (Figure 1C). The Result module provides the visualization of segmented cellular morphologies while showing all statistical information as well. Particularly, the segmented 3D embryo is shown with each cell labeled in different colors (Figure 1D). Users can apply quality control to statistical data to filter out confident results. The Train module allows users to retrain the model based on a specific dataset to deliver optimal results.