Scripts for vacuole quantification in 3D whole-mount 2-photon images of D. melanogaster brains
Main publication: https://www.cell.com/iscience/fulltext/S2589-0042(23)02129-6
Main publication Citation : Elguero, J. E., Liu, G., Tiemeyer, K., Bandyadka, S., Gandevia, H., Duro, L., Yan, Z., & McCall, K. (2023). Defective phagocytosis leads to neurodegeneration through systemic increased innate immune signaling. iScience, 0(0). https://doi.org/10.1016/j.isci.2023.108052
Protocol paper: https://www.sciencedirect.com/science/article/pii/S2666166724001825?ref=cra_js_challenge&fr=RR-1
Protocol paper citation: Liu G, Bandyadka S, McCall K. Protocol to analyze 3D neurodegenerative vacuoles in Drosophila melanogaster. STAR Protoc. 2024;5: 103017. doi:10.1016/j.xpro.2024.103017
To identify whether there is a statistically significant difference in the number, size, and spatial distribution of neurodegeneration vacuoles in whole-mount 2-photon microscopy images of D.melanogaster brains
Whole-mount 2-photon imaging
- Red : phalloidin (muscle)
- Green : Hoechst (nucleus)
Z-stacks acquired from confocal microscopes need to be converted to RGB stacks or 8bit greyscale stacks in the TIFF format for use with WEBKNOSSOS for annotation.
The FIJI macro provided - code/maketiff.ijm
can be used to batch process raw images.
The macro takes in .nd2
files and performs the following operations for each .nd2
file -
- split channels
- merge channels
- convert to RGB tiff
- convert to 8bit greyscale tiff
The macro can be modified to work with other common microscopy raw formats such as .oif
When the macro is run in FIJI, it will open a dialog box to select the input folder containing all the .nd2
files. Once the input folder is selected, another dialog box to get the output destination folder will open up. Once the input and output folder are selected, all the .nd2
files in the input folder will be converted to RGB stacks and 8-bit greyscale stacks and will be available in the selected destination folder for downstream analysis.
Make sure to not include any other files other than the necessary .nd2
files in the input folder selected.
Input .nd2
files must not contain any spaces or special characters other than underscores in the filename. An example of an ideal input filename is - best_input_file_123.tif
- Use prairie reader plugin to import xml files into FIJI
- Convert stack to RGB
- Save merged RGB z-stacks in Tiff format
- Upload RGB TIFF z-stacks of individual brains to Webknossos
- Create a new volume annotation
- Identify each vacuole visually. For each vacuole
- create a new segment id
- Paint over vacuoles in some z-stack levels using the quick select automatic segmentation tool
- Use the interpolation tool to automatically fill in the vacuole
- When the vacuole is fully annotated, create a 3d mesh for the segment
- Repeat the above steps to annotate the whole brain
- Download the mesh for vacuoles and whole brain as stl files
Example annotated data is deposited at Zenodo:
# Create a Python3 virtual environment and activate
python3 -m venv vacuolequant
source vacuolequant/bin/activate
# Install all packages at once
pip3 install -r requirements.txt
# Alternatively install necessary packages individually
pip3 install ipykernel
pip3 install numpy-stl
pip3 install matplotlib
pip3 install seaborn
pip3 install numpy
pip3 install pandas
pip3 install jupyter
pip3 install scipy
# Create Jupyter notebook kernel
python3 -m ipykernel install --name=vacuolequant --user
jupyter notebook
# Run notebook for vacuole quantification, statistical analysis, and visualization
quantify_vacuoles_fromMesh_LiuBandyadka2024.ipynb # To quantify vacuoles from mesh files
quantify_vacuoles_fromWebknossosexport_LiuBandyadka2024.ipynb # To quantify vacuoles from csv files