The Francis Crick Institute
Browse

An image processing and analysis methodology for investigating the reshaping of the epithelium in the zebrafish otic vesicle

Download (1.48 MB)

Poster presented as part of the Crick BioImage Analysis Symposium 2023.

The zebrafish inner ear presents an optical challenge due to its mobile and densely packed cells, presenting issues for cell segmentation and tracking during development. The depth of the otic vesicle in the zebrafish embryo presents light scattering issues, and the technical compromises involved in imaging a large structure like the otic vesicle may reduce frame rates or stack size in the interest of maximising storage space, producing challenges in tracking cells. Tracking can be implemented out of segmenting and tracking cell membrane markers, or can follow a more targeted approach of following cell nuclei. Cell membrane segmentation presents challenges in the form of merging cells in the z plane where z stacks are not of sufficient resolution, or in the x/y plane when the fluorophore distribution through the membrane is uneven. This uneven distribution of membrane fluorophore may be compounded by optical anisotropy if imaging conditions, due to depth within tissue or uneven illumination. We present our methodology to probe the development of the otic epithelium. We produced a script which chains image preprocessing steps: noise reduction, rough thresholding and 3D erosion, used to pinpoint locations of nuclei in optically difficult regions of epithelium. This step discerns closely associated structures which may get grouped together using standard watershed based segmentation techniques. This methodology presents a strategy to process and analyse dynamic and optically anisotropic structures from in vivo data.


Permission has been given by authors to upload to Crick Figshare. Copyright remains with the original authors.

History

Usage metrics

    The Francis Crick Institute

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC