DANEELpath: Digital image Analysis for NEuroblastoma modELs

DANEELpath is an open-source image analysis toolkit developed to work with neuroblastoma models.

At the moment DANEELpath tools are focused on obtaining data from 3D cell culture hydrogels histological slides. Nevertheless the tools can be used to work with in vivo and human histological samples.

DANEELpath is built to work inside QuPath, a popular software for digital pathology.

Our YouTube tutorial showing how to install DANEELpath is available here:YouTube Playlist

The tutorials can be downloaded tooguether with the DL model weights from INCLIVA cloud here

The original preprint, with all the details can be found here: DANEELpath: Open-source Digital analysis tools for histopathological research. Applications in NEuroblastoma modELs. bioRxiv 2025.03.31.646267; doi: https://doi.org/10.1101/2025.03.31.646267

What can DANEELpath do?

The neuroblastoma cells grow in cell cluster fashion inside of 3D hydrogels. Our focus is to study of the effect glicoprotein Vitronectin in the growth and progression of neuroblastoma cells.

When we do sections of the Hydrogels and do an H&E staining, we can see the clusters in a huge range of sizes. When we do an immunostaining against Vitronectin, we observe a peri-cluster deposit , that we designated as Vitronectin-Corona. However, when we grow the cells in matrices in wich we added Vitronectin, when we do the immunostaining agains Vitronectin, a faint peri-cluster intensity appears, that we designed pale-halo.

DANEELpath tools have been designed to streamline the process of segmentation of cell clusters, peri-cluster structures. Furtheremore, we designed tools to obtain morphological and spatial features from cell clusters. Those tools include:

1. Segmentation tools:

• Semantic Segmentation of cell clusters using a U-Net for H&E and Vitronectin stainings
• Pipeline for segmentation and feature extraction of peri-cluster structures.

2. Spatial feautures tools:

• Determination of center and periphery regions with equal area of a 2D shape
• Descriptors of the distances between cells nearest neighbors inside cell clusters
• Count the number of neighbors arround cell clusters, with three neighborhood criteria

Nice but... Do I need programming experience to use DANEELpath?

You do NOT need programming experience to use DANEELpath!

We are aware that not all the users have programming experience, and we do not want this to limit DANEELpath use. In the DANEELpath_GUI folder, you will find a set of groovy scripts that, when run in QuPath, will create a Graphical User Interface to interact.

Nevertheless, we also want to support users with more experience. In the DANEELpath_Scripts folder you can find the groovy script version of the tools, with the same parameters as in the GUI, which enables to run the scripts for multiple images in QuPath.

To facilitate DANEELpath use, we created a YouTube tutorial showing the installation process. Furthermore we provide a series of tutorials showing how to use all the tools

INSTALLATION

Here you will find the instructions for DANEELpath setup. If you have any problems, you can check our YouTube videos.

1. Install QuPath and Cellpose extension

1.1 QuPath

DANEELpath works inv QuPath version 0.5 and above. You can download QuPath from this link: https://qupath.github.io/

If you have problems during QuPath installation, you can check this link: https://qupath.readthedocs.io/en/0.5/docs/intro/installation.html

1.2 Install qupath-extension-cellpose

DANELpath uses a functionality inside the incredible qupath-extension-cellpose from BIOP.

DANEELpath need the version 0.9.6 or above of the BIOP extension. You can download the last realease of the extension in this link: https://github.com/BIOP/qupath-extension-cellpose/releases

Once download, unzip the folder and drag & drop qupath-extension-cellpose-[version].jar onto the main QuPath window. Then restart QuPath (not the computer)

NOTE: You do not need to install Cellpose itself, just the BIOP extension .jar file. However, we encorage you to explore the incredible qupath-extension-cellpose.

2. Download the Github Repository, the U-Net checkpoints and the tutorials.

2.1 GitHub Repository

You can download the GitHub repo as zip folder and extract the content in the desired folder in you machine.

If you have Git, you can create your destiny folder in your machine, open a Git Bash terminal and run the command:

git clone https://github.com/iviecomarti/DANEELpath.git

By doing this, you will have all the files present in this repository. However we still need to download the weights of the models and the tutorials

2.2 U-Net checkpoints and tutorials.

You can download the U-Net weights and tutorials from INCLIVA cloud here: DANEELpath_Unets_and_Tutorials

Once downloaded, unzip the folder. Inside you will find two folders: U-Net_checkpoints and Tutorials

IMPORTANT: DO NOT CHANGE THE FOLDERS NAME

You need to place, at least the folder U-Net_checkpoints to the same folder as all de files of the GitHub Repository have been downloaded. You should have somehting like this:

Now that we downloaded the models, we need to do the set-up to install Miniconda and PyTorch.

3. Download Miniconda and setup the PyTorch enviroment

3.1 Miniconda

Install Miniconda for you OS. You can find the instructions to install it here: https://docs.anaconda.com/miniconda/install/#

If you have any questions, please check our YouTube video.

NOTE: We will show the GPU support in section 5

3.2 Install the enviroment and check torch installation

In DANEELpath folder, you will see the enviroments folder. Inside that folder you will find DANEELpath_env_[OS]_CPU.yml. You have the .yml files for Windows, Linux and AppleSilicon

Open a Miniconda terminal and go to the enviroments folder where you unziped DANELpath.

If you are in Windows11 you need to run the following command in the Anaconda terminal:

conda env create -f DANEELpath_env_Windows_CPU.yml

This will take a little bit of time to install all the dependencies.

When the installation finishes, we neeed to check if torch is installed succesfully. Open the DANEELpath enviroment by typing in the Anaconda terminal:

conda activate DANEELpath

Then run this line of code:

python -c "import torch; x = torch.rand(5, 3); print(x)"

You will see something like this in the terminal if everything goes okay:

tensor([[0.3380, 0.3845, 0.3217],
        [0.8337, 0.9050, 0.2650],
        [0.2979, 0.7141, 0.9069],
        [0.1449, 0.1132, 0.1375],
        [0.4675, 0.3947, 0.1426]])

3.3 Obtain the Python path

If you are in a Windows machine run in the terminal:

where python

You will obtain an output simmilar to this one if you are in Windows11:

C:\Users\User\anaconda3\envs\DANEELpath\python.exe

If you are in a Linux/Mac run in the terminal:

which python

You will obtain an output simmilar to this one in Linux/Mac

/home/User/miniconda3/envs/DANEELpath/bin/python

Copy the output for DANEELpath python.exe, and it is IMPORTANT to change the direction of the bars from \ to / if you are in Windows11. The above example you need to do: C:/Users/User/anaconda3/envs/DANEELpath/python.exe

For Linux and Mac, the bars are in the correct direction. Furthermore, you do not have python.exe after asking the path. You should copy just: /home/User/miniconda3/envs/DANEELpath/bin/python

4. Set the Python Path in Groovy scripts for U-Net

To be able to use the U-Nets, you will need to do a final adjustment in the semantic segmentation Scripts.

4.1 Change in script for DANEELpath GUI

Go to DANELpath > DANEELpath_GUI > 1.GUI_U-Net_Semantic_Segmentation. Open the script in QuPath and you will see this piece of code at the beggining of the script:

//Python conda settings
def pythonElements() {
    
    def daneelFolder = "path/DANEELpath/folder"
    def pathPythonEnviromentEXE = "path/python.exe"
    
    return [modelsFolder:danelFolder,exePath:pathPythonEnviromentEXE]
    
}

In daneelFolder put the path where you downloaded DANEELpath. IMPORTANT: you must put the separators in this direction / In pathPythonEnviromentEXE put the python.exe path that you obtained in 3.3. IMPORTANT: you must put the separators in this direction /

Save the groovy script.

4.1 Change in script for DANEELpath Script

Go to DANELpath > DANEELpath_Scripts > 1.Semantic_Segmentation > 1.1.Run_U-Net_Script. Open the script in QuPath and you will see this piece of code at the beggining of the script:

//Python conda settings
def danelFolder = "path/DANEELpath/folder"
def pathPythonEnviromentEXE = "path/python.exe"

In daneelFolder put the path where you downloaded DANEELpath. IMPORTANT: you must put the separators in this direction / In pathPythonEnviromentEXE put the python.exe path that you obtained in 3.3. IMPORTANT: you must put the separators in this direction /

Save the groovy script.

Now you can use DANEELpath!

5.OPTIONAL: NVIDIA GPU for Windows11 and Linux

We can take advantage of NVIDIA GPU acceleration for inference in DANEELpath. The U-Net fits in a 4GB VRAM GPU, so there is no need of huge investment. However to achieve this we need to do the setup for CUDA.

5.1 Download the NVIDIA drivers for you GPU

If you consider to use GPU, probably you have already done this step before. But just in case, make sure you have your GPU drivers updated.

You can find your the drivers for the GPU in this web: https://www.nvidia.com/en-us/drivers/

Put the characteristics of the GPU, download, and install the drivers

5.2 Install PyTorch with CUDA.

The enviroment for DANEELpath installs torch without CUDA support. The easiest way to achieve the GPU support is to uninstall this version of torch and install the version with GPU. Follow the steps:

1. activate the enviroment:

conda activate DANEELpath

2. Remove torch CPU:

pip uninstall torch

3. Install torch GPU. Here we tested CUDA 11.8 and works with Python 3.10.

pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

4. Check if PyTorch detects GPU by running this command

python -c "import torch; print(f'CUDA Available: {torch.cuda.is_available()}, Devices: {[torch.cuda.get_device_name(i) for i in range(torch.cuda.device_count())] if torch.cuda.is_available() else None}')"

You should obtain a result like this one:

CUDA Available: True, Devices: ['NVIDIA GeForce RTX 4070 Ti SUPER']

CITATION

If you use DANEELpath, you need to cite the following references:

Bankhead, P. et al. QuPath: Open source software for digital pathology image analysis. Scientific Reports (2017). https://doi.org/10.1038/s41598-017-17204-5

Burri O, Sobolewski P, Fehlmann T. BIOP:qupath-extension-cellpose. 2024 Available from: https://zenodo.org/records/13752036

And also our reference

Vieco-Martí et al. DANEELpath: Open-source Digital analysis tools for histopathological research. Applications in NEuroblastoma modELs. bioRxiv 2025.03.31.646267; doi: https://doi.org/10.1101/2025.03.31.646267

ISSUES

If you find any problems using DANEELpath please open an issue. We will look at it as soon as possible 😊