posterior_neuruloid_patterning_segmentation

Analysing posterior neuruloid patterning using segmentation method to obtain intensity trends.

Suggested set-up of virtual environment for project

Create virtual environment for CellPose https://www.cellpose.org that works for your device. Other libraries are standard, listed in requirements.txt.

conda activate cellpose
conda config --env --add channels conda-forge
conda install --file requirements.txt

Pipeline

Details on usage of folders and files. Use notebooks to run pipeline (some are optional) rougly in the order listed. More details of the pipeline are in notebook files.

neuruloid_2D_segmentation/
├── fiji_macros/ 
        ├── get_dapi_only.ijm # extract dapi channel only for segmentation
        ├── clahe_images.ijm  # correct stitched dapi images in a folder
├── neuruloid_segmentation/ # contains .py files of all the functions used in notebooks
├── notebooks/ # run processing in notebooks
        ├── choose_best_2D_slice.ipynb      # choose most focussed slice from 3D stack
        ├── correcting_stitching.ipynb      # take clahe dapi images (saved from fiji) and combine with other channels
        ├── 2D_segment_24hr.ipynb           # Segment image, save masks and save raw data to "intensity_data" - ensure dapi channels are extracted into a separate "dapi/" folder
        ├── plotting_mean_intensities.ipynb # Process raw dataframe by getting mean of intensities depending on distance from edge and save to "../processed_data"
        ├── fitting_rd_with_tbxt.ipynb      # comparing data from "../processed_data" to rd simulations
        ├── quantifying_segmentation_accuracy.ipynb # calculate sensitivity and precision of segmentation
        ├── unconstrained_neuruloid.ipynb # analyse TBXT clusters from unconstrained neuruloid
        └── import_helper.py # to import scripts from "../neuruloid_segmentation"

Data storage

Git ignores folder 'images/' in this repo. Store images in folder named 'images/' with the following data structure to run the notebook (or however you like just change in notebook):

neuruloid_2D_segmentation/
├── ...
├── images/                           # this folder is ignored by git since will be too large
    └── dataset_name/                 # name of dataset (e.g., single_z_slices_24hr)
        └── experiment type/          # data for experiment type (e.g., circle_ctrl)
            └── neuruloid_parameter/  # data for some neuruloid parameters (e.g., ctrl_350)
    │           ├── sample_1.tif      # full tiff images with all channels
                ├── sample_2.tif
                ├── ...
                └── dapi/             # extract only dapi images from the full tiff images for segmentation
                    ├── sample_1.tif
                    ├── sample_2.tif
                    └── ...
                └── masks/            # masks after segmentation
                    ├── sample_1.npy
                    ├── sample_2.npy
                    └── ...
├── intensity_data/                   # this folder contains extracted data for each neuruloid 
    └── dataset_name/                 # name of dataset (e.g., single_z_slices_24hr)
        └── experiment type/          # data for experiment type (e.g., circle_ctrl)
            └── neuruloid_parameter/  # data for some neuruloid parameters (e.g., ctrl_350)
                ├── sample_1.csv      # csv file for extracted data for each nuclei
                ├── sample_2.csv
                └── ...
        └── pixel_size.txt            # dictionary containing pixel sizes 
├── processed_data/                   # this folder contains processed data across multiple organodis and normalised 
    └── dataset_name/                 # name of dataset (e.g., single_z_slices_24hr)
        └── experiment type/          # data for experiment type (e.g., circle_ctrl)
            └── parameter_1.csv       # data for some neuruloid parameters (e.g., ctrl_350.csv)
            ├──  parameter_2.csv  
            └── ...
        └── pixel_size.txt            # dictionary containing pixel sizes 
    └── simulation_intensity.csv      # data from reaction-diffusion simulation for gamma parameters
    ```