Software and tools to perform Vasculature and DAPI analysis on mouse retina
This is a submodule that's part of a larger project to perform retinal analysis. Please stay tuned for more retinal analysis with microglia and astrocyte quantifications.
The susequent DAPI and Vasculature analyses, expects the data be of the following folder structure. For each sample, there should be a folder with two subfolder called , raw and isotropic. Inside isotropic we need three files, DAPI stack, DAPI stack in XZ cross sectional view and blood vessel stack. A Sample structure would look like this
If you have individual DAPI stained images or Iba1 stained images then use the data_preparation_single.ijm macro file to prepare your data. If you have a multichannel image, with series blocks then use the data_preparation_batch.ijm macro file. These macro files will automatically create a folder structure of the required format.
Once the data is prepared, the DAPI pipeline can be exeuted using the main_pipeline_DAPI.ipynb python notebook inside the 2_DAPI_Analysis folder. The code will automatically download the neural network model weights and perform segmentation and analysis.
Note : Input and output path inside the notebook should be the parent folder which will contain mouse sample names inside. For example, the "prepared_data" folder can be an input folder.
The installed python environment from the yml file comes with a jupyter environment. Below is one of the ways you can execute the notebook,
- Open jupyter lab after navigating to the folder
cd Retinal_Analysis_Vas
jupyter lab
- Navigate to the python notebook file inside jupyter lab and execute the notebook
For each sample, there will be a "C4-DAPI-XZ_reconstructed_cleaned_xy.tif". This is the final segmentation output file. The output segmentation will be an instance mask delineating the layers - RNFL, GCL, IPL, INL, OPL, ONL, and IS/OS.
The 3D sample output will look like below,
To visualize in 3d, you can use napari which comes along with the installed environment.
After performing DAPI analysis, vasculature analysis can be performed in two steps. First step is the segmentation step which can be completed by executing the 1_Blood_Cessel_Segmentation.ipynb. The second step is to perform a feature extraction and branch analysis on the segmented image which can be done by executing the 2_Blood_Vessel_Post_Segmentation.ipynb.
Note : Input and output path inside both the notebooks should be the parent folder which will contain mouse sample names inside. For example, the prepared_data folder in the previous folder structure image can be an input folder.
The final instance segmentation output file is "C1-Icam2-Blood-Vessels_branch_labels_nnunet.tif". The final sementic segmentation output file is "C1-Icam2-Blood-Vessels_nnunet_reconstructed_cleaned_filtered.tif". For each sample csv feature files will be generated.
A 3D segmentation output will look like below,
