Axoden

Manuscript can be found on eNeuro (pdf)

Table of Contents

• How to use AxoDen
  • Streamlit App
  • Standalone GUI
• Installation
  • With pip
  • Git clone
• Using AxoDen in your code

Overview of AxoDen

AxoDen is an open source platform that facilitates and streamlines the quantification of fluorescently labeled axonal projections of a given brain area.

AxoDen accepts .tif images, transforms them into gray scale and uses the Otsu method to set a threshold that distinguishes signal from background. Because it collects the pixel values of any color channel which sum is above zero, users can use any fluorophore and obtain the same quantification accuracy. In addition, given that the input are images of projections, AxoDen can be applied to virtually any animal species.

AxoDen is a plug an play tool, designed for users with no coding background as well as for developers who wish to integrate the AxoDen functions into their own scripts (read the section "How To Use AxoDen").

How To Use AxoDen

The simplest way to use AxoDen is to simply use the deployed web app at axoden.streamlit.app. It will allow you to upload images directly through a browser and download the results. The files that you upload are only ever kept in memory and discarded once you disconnect from the web app. And we never store or collect anything you upload.

However, if you do not want to upload the images or need to process large amounts of images, we provide a standalone application with a GUI that you can download from here.

Streamlit App

Whether you use the deployed app axoden.streamlit.app or run it locally (see Installation), the easiest way to get started is to check the Tutorial & How To page that you see on the left pane when the app starts. This should explain the steps required and also give you a few tips on how to prepare the images before you upload them.

Standalone GUI

If you download the standalone gui from our release page, there are a few considerations:

Preparing folder and images

Axoden processes folders, not single files. So all the files that you want to process should be in a folder you prepare.

Naming Convention

• All images need to be in .tif format.
• The filename of the images matter! Rename the files to the following format: subjectID_brainregion_group_WhateverYouWant.tif

Tip

• Do not use spaces in the name of the file
• The underscore ("_") character is used to separate the subjectID, brainregion and group
• At least subjectID and brainregion are required, e.g. "animal1_ACC.tif" is valid, "myfile.tif" does not work
• The group is optional, but can be helpful if you want to further process the output from Axoden, e.g. if you want to do your own statistics.

Image Masking

If you want to mask your images to analyze a precise brain region:

1. Use Adobe Illustrator or Power Point to overlay the brain atlas onto your image and set the boundaries of your brain region of interest.
2. Download ImageJ or FIJI if you do not have it installed.
3. Open the image of interest.
4. Select the freehand selection tool on the top menu.
5. Draw the boundaries of the region of interest according to the overlay you created in step 1. On the top menu select in order:
   • Edit > Clear Outside
   • Image > Crop
6. If the signal to noise ratio needs to be enhanced:
   • On the top menu select Image > Adjust > Brightness > Crop
   • Decrease the "Maximum" value using the sliding bar to increase fluorescence intensity.
   • Increase the "Minimum" value using the sliding bar to decrease background fluorescence
7. Save the resulting image following the AxoDen naming conventions.

Tip

Use the histogram (Analyze > Histogram) to confirm that the background fluorescence in the brain region does not contain zero values.

Running the GUI

Execute the application, and you should be presented with the follwoing GUI:

• Select the folder you prepared in the previous step. Use the Select Folder button to open a folder selection dialog
• The output folder will be set to the subfolder Axoden_output but you can change it. If the folder does not exist, it will be created.
• Make sure the pixel size is set to the correct value. This depends on both the objective used and your device!
• If you use masked images as described in step Image Masking, leave the check as it is. If you did not mask the images and want to process original, unmasked images, remove the check.
• By default, AxoDen will collect all data, then plot the summary and finally the summary axis data.
• Finally click Run Volume Projections to start the processing.
• You will find the results in the Output Folder.

Installation

Note

This section applies only to people who want to use AxoDen in their python code, or want to run AxoDen from code.

There are two ways to install axoden: Through pip (recommended) or by cloning the repo from github.

Pip Install

For most users, this is the preferred option. Install axoden through pip:

  pip install axoden

If you also want the extra gui dependencies, specify it in the pip install:

  pip install axoden[gui]

Running the GUIs

Assuming you installed the gui dependencies with the above pip install axoden[gui], you can start the tkinter with

  python -m axoden.gui.gui_projections_quantification

If you get an error saying TkInter is not installed, see the Install TkInter section.

To start the streamlit app locally, run:

  python -m axoden.gui.streamlit_app

This is the same interface that is running as the Streamlit Web App, but when running it locally you will benefit from faster processing time, and uploads should be almost instantaneous.

Clone from github

If you want to make changes to axoden, you can clone the full code from the github repo.

Clone the repo

  git clone https://github.com/raqueladaia/AxoDen.git

We suggest to set up a virual environment. For more information regarding virtual environments, see https://docs.python.org/3/library/venv.html

Install the dependencies

Use pip to install the dependencies (typically in your virtual environment):

pip install -r requirements.txt

This will install all required python dependencies to fully use axoden, including running the streamlit app locally. The tkinter library cannot be installed through pip and has to be installed separately.

There are two other requirements files that you might want to use instad of the above.

• requirements_no_gui.txt: Excludes streamlit, but allows you to use axoden as a library, as well as use the tkinter gui.
• requirements_dev.txt: For development purposes, includes everything including testing and linting (tox, pytest, black, ruff).

Install TkInter

For windows, tkinter is installed when installing python. If you get an error complaining the module named 'tkinter' is not found when executing the python code, you will have to modify the python installation. For that, launch the python installation again and in the installer, select 'Modify'. Make sure that the checkbox 'tcl/tk and IDLE' is checked!

For linux and Mac, you can install tkinter via command line:

# Linux
sudo apt install python3-tk
# Or for a specific version of python, e.g. python 3.11:
sudo apt install python3.11-tk

# Mac (not tested):
brew install python-tk

Running the GUI

Simply execute the file axoden/gui/gui_projections_quantification.py, e.g. through a terminal:

  python axoden/gui/gui_projections_quantification.py

Running Streamlit App Locally

You can run the streamlit app locally with

  streamlit run axoden/gui/streamlit_app/1_📊️_Axoden.py

This should automatically open a browser window to the app. If not, streamlit will post the URL where you can reach the web app, usually at http://localhost:8501.

This is the same interface as the Streamlit Web App.

Using axoden in your code

Below you find a few examples of how to use axoden. You can see the full documentation here.

You can process a single image. This will not save any results or plots.

import axoden

img_path = "test_images/745_TH-CL.tif"
fig, data, data_axis = axoden.process_image(img_path, is_masked=True, pixel_size=0.75521)
fig.show()

Or you can process all images in a folder and save the results, similar to what the gui does:

import axoden

input_folder = "test_images"
output_folder = "results"
df_data, df_data_axis = axoden.process_folder(input_folder, pixel_size=0.75521, is_masked=True, output_folder=output_folder)
axoden.write_summary_data_plot(output_folder, df_data, project_name="my project")
axoden.write_signal_intensity_along_axis_plot(output_folder, df_data_axis, pixel_size=0.75521, project_name="my_project")

It is also possible to not save any output, this might be useful if you further process the results.

import axoden

input_folder = "test_images"
df_data, df_data_axis = axoden.process_folder(input_folder, pixel_size=0.75521, is_masked=True, save=False)

output_folder = "results"
axoden.save_table(df_data, output_folder, "projections_quantifications.csv")
axoden.save_table(df_data_axis, output_folder, "projections_quantification_along_axis.csv")

fig_data = axoden.plot_summary_data(df_data, "my_project")
fig_data_axis = axoden.plot_signal_intensity_along_axis("my_project", df_data_axis, pixel_size=0.75521)

fig_data.show()
fig_data_axis.show()